JP7755028B2 - Control device, cleaning device, cleaning device control method, and control program - Google Patents

Description

The present invention relates to a control device that controls the movement of a cleaning device that cleans pipes.

Boilers used to generate electricity using waste heat from incinerators are equipped with numerous pipes. These pipes are located in the exhaust path of the exhaust gas generated by the incinerator, and fluids such as water and steam flow through them. For this reason, these pipes are also called boiler water pipes. The thermal energy generated by incineration is recovered through heat exchange between the exhaust gas and the fluid flowing through the boiler water pipes.

As boiler water tubes are exposed to exhaust gases, fly ash generated in incinerators and other materials adhere to and accumulate on the surface of the boiler water tubes, reducing their heat exchange efficiency. Therefore, the surface of the boiler water tubes must be cleaned regularly. However, cleaning boiler water tubes by hand is not easy, and automatic cleaning using cleaning devices has been considered for some time.

For example, Patent Document 1 below discloses a cleaning device equipped with a water tube group traveling cleaning device that is equipped with a cleaning tool and moves in the axial direction of a boiler water tube group. This water tube group traveling cleaning device is lowered between the boiler water tubes to be cleaned and moves along the side of the boiler water tubes to clean. Because the water tube group traveling cleaning device cannot move in the tube row direction, it is moved while housed in a tube row direction moving device that moves in the tube row direction of the boiler water tube group, and then lowered between the boiler water tubes to be cleaned next.

Japanese Patent Application Laid-Open No. 2006-138572

The cleaning device in Patent Document 1 can only move along the boiler water tubes when the water tube group traveling cleaning device is lowered between the boiler water tubes. Therefore, when starting cleaning, the water tube group traveling cleaning device must first be stopped in a position where it can be lowered between the boiler water tubes.

However, it is difficult to always stop a water tube group traveling cleaning device in the correct position. For this reason, the cleaning device in Patent Document 1 is equipped with a stop position correction arm for correcting the stopping position of the water tube group traveling cleaning device.

As such, the cleaning device of Patent Document 1 lacks versatility in that it requires a special mechanism, such as a stop position adjustment arm, to automatically move the cleaning device. One aspect of the present invention aims to achieve general-purpose movement control for cleaning devices that clean pipes.

In order to solve the above problem, one aspect of the present invention provides a control device that cleans multiple pipes arranged in parallel. The control device includes a deviation calculation unit that calculates the deviation of the cleaning device from a predetermined reference position based on the detection values of a pair of proximity sensors that are attached to symmetrical positions on the left and right sides of the cleaning device and detect the pipes, and a movement control unit that controls the movement of the cleaning device based on the deviation. The pair of proximity sensors are positioned so that when one proximity sensor is located directly above the pipe, the other proximity sensor is located at the detection limit of the pipe adjacent to the pair of proximity sensors, and the deviation calculation unit calculates the deviation using an approximation equation that approximates the relationship between the difference in detection values of the pair of proximity sensors and the deviation.

In order to solve the above-mentioned problems, one aspect of the present invention provides a cleaning device control method executed by a control device, which includes: a deviation amount calculation step of calculating the amount of deviation of the cleaning device from a predetermined reference position based on detection values of a pair of proximity sensors attached to symmetrical positions on the cleaning device that cleans multiple pipes arranged in parallel, and detecting the pipes; and a movement control step of controlling the movement of the cleaning device based on the deviation amount, wherein the pair of proximity sensors are positioned so that when one proximity sensor is positioned directly above the pipe, the other proximity sensor is positioned at the detection limit of the pipe adjacent to the pair of proximity sensors, and the deviation amount calculation step calculates the deviation amount using an approximation equation that approximates the relationship between the difference in detection values of the pair of proximity sensors and the deviation amount.

One aspect of the present invention makes it possible to achieve general-purpose movement control of a pipe cleaning device.

1 is a block diagram showing an example of a configuration of a main part of a control device according to a first embodiment of the present invention; FIG. 1 is a diagram showing an overview of a cleaning system including the control device. 10 is a diagram showing an outline of pipe detection and inclination angle identification by the control device. FIG. 10A and 10B are diagrams illustrating a specific example of pipe detection by the control device. 4 is a diagram illustrating an outline of a method for calculating a deviation amount by a deviation amount calculation unit included in the control device. FIG. 10A and 10B are diagrams illustrating the relationship between the positional relationship between a proximity sensor and a pipe and the detection value of the proximity sensor. 10A and 10B are diagrams illustrating the relationship between the positional relationship between a proximity sensor and a pipe and an approximation formula for a detection value of the proximity sensor. FIG. 10 is a diagram showing a function for calculating the amount of deviation from the difference between the detection values of the proximity sensors. 10A and 10B are diagrams illustrating an example of the operation of the cleaning device during centering. 10A and 10B are diagrams illustrating a method for calculating the travel time of the cleaning device during centering. 10 is a flowchart illustrating an example of a method for controlling the cleaning device based on the tilt angle. 10 is a flowchart illustrating an example of a method for controlling the cleaning device based on the amount of deviation. 10A and 10B are diagrams illustrating an example of calculating a deviation amount according to the second embodiment of the present invention.

[Embodiment 1]
(Cleaning system overview)
An overview of a cleaning system 5 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a diagram showing an overview of the cleaning system 5. The cleaning system 5 is a system for cleaning the surface of a pipe PI and includes a control device 1 and a cleaning device 2. The control device 1 is a device that controls the operation of the cleaning device 2. The cleaning device 2 is a cleaning device for the pipe PI that operates under the control of the control device 1. The upper part of Fig. 2 shows a perspective view of the cleaning device 2, and the lower part of the same figure shows a side view of the cleaning device 2.

The pipes PI are straight pipes arranged in multiple rows at equal intervals in the horizontal direction. The pipes PI are also arranged in multiple rows in the vertical direction. In this embodiment, an example is described in which the pipes PI are the piping of a boiler (not shown) used to generate electricity using waste heat from an incinerator, i.e., the boiler water pipes mentioned above. Of course, the cleaning system 5 can also be used to clean pipes other than boiler water pipes, as long as they are strong enough for the cleaning device 2 to travel over them and at least a portion of them is straight.

As shown in the figure, the cleaning device 2 comprises a main body 21, a crawler 22, a hose reel 23, and a hose 24, and a pantograph 25 is housed inside the main body 21. Furthermore, as shown in the side view of Figure 2, a water outlet 26 is provided at the tip of the pantograph 25. Furthermore, a housing 27 is provided at the front end of the main body 21.

Although only the left crawler 22 is shown in Figure 2, a pair of crawlers 22 are provided at symmetrical positions on either side of the main body 21. The crawlers 22 function as a travel device that moves the cleaning device 2 forward, backward, and rotates it in a horizontal plane. Other types of travel devices, such as wheels, may be used instead of the crawlers 22.

The hose reel 23 is used to wind the hose 24, which is a pipe that delivers cleaning liquid (e.g., water) for the pipe PI to the water outlet 26. The pantograph 25 is retractable using a link mechanism, allowing it to extend downward from the main body 21 during cleaning and retract during movement. The water outlet 26 is also connected to the hose 24, and discharges the cleaning liquid delivered from the hose 24 to the side.

In other words, the cleaning device 2 extends the pantograph 25, positions the water outlet 26 to the side of the pipe PI to be cleaned, and in this state, sends cleaning liquid from the hose 24 into the water outlet 26, thereby cleaning the pipe PI with water pressure. The method for cleaning the pipe PI is optional, and for example, a cleaning tool such as a brush may be pressed against the pipe PI to clean it.

The cleaning device 2 lowers its pantograph 25 between the pipes PI to be cleaned to perform cleaning, so it must move between the pipes PI as a preliminary step to cleaning. To perform this movement automatically, a camera and proximity sensor are installed inside the storage section 27 of the cleaning device 2 (neither is shown in Figure 2).

The control device 1 then acquires the image captured by the camera device and the detection value of the proximity sensor, and controls the cleaning device 2 based on these. This allows the cleaning device 2 to automatically move between the pipes PI to be cleaned and have the cleaning device 2 clean the pipes PI.

(Configuration of control device)
The configuration of the control device 1 will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of the configuration of the main parts of the control device 1. As shown in the figure, the control device 1 includes a control unit 10 that controls each part of the control device 1 in an integrated manner, and a storage unit 11 that stores various data used by the control device 1. The control device 1 also includes a communication unit 12 that allows the control device 1 to communicate with other devices, an input unit 13 that accepts various data input to the control device 1, and an output unit 14 that allows the control device 1 to output various data. The control unit 10 includes a pipe detection unit 101, an angle identification unit 102, a deviation amount calculation unit 103, and a movement control unit 104.

The pipe detection unit 101 detects the pipe PI from images captured by a camera attached to the cleaning device 2. The images may be acquired via the communication unit 12 or the input unit 13. The angle identification unit 102 then identifies the inclination angle of the pipe PI detected by the pipe detection unit 101.

The deviation amount calculation unit 103 calculates the amount of deviation of the cleaning device 2 from a predetermined reference position based on the detection values of a pair of proximity sensors that detect the pipes PI and are attached to symmetrical positions on the left and right of the cleaning device 2. In this embodiment, the reference position is the center position of two pipes PI arranged adjacent to each other in parallel (a position equidistant from the two pipes PI), and an example is described in which the deviation amount calculation unit 103 calculates the amount of deviation between the center position in the left-right direction of the cleaning device 2 and the above-mentioned center position.

Of course, the reference position can be set as appropriate and is not limited to this example. For example, the center position of the pipe PI can be set as the reference position, and the amount of deviation between the center position of the cleaning device 2 in the left-right direction and the center position of the pipe PI can be calculated.

The movement control unit 104 controls the movement of the cleaning device 2 based on either or both of the tilt angle identified by the angle identification unit 102 and the amount of deviation calculated by the deviation amount calculation unit 103. As will be described in detail below, the movement control unit 104 performs, for example, rotation control to change the orientation of the cleaning device 2 and control to move the cleaning device 2 to the center position of the adjacent pipe PI. Note that here, rotation control that does not change the position of the control device 2 is also included in the category of movement control.

As described above, the control device 1 includes a pipe detection unit 101 that detects the pipe PI from images captured by an imaging device attached to the cleaning device 2 that cleans the surface of the pipe PI, an angle identification unit 102 that identifies the inclination angle of the pipe PI detected by the pipe detection unit 101, and a movement control unit 104 that controls the movement of the cleaning device 2 based on the identified inclination angle.

The tilt angle of the pipe PI shown in the image captured by the camera attached to the cleaning device 2 reflects the orientation of the cleaning device 2 relative to the pipe PI. If the orientation of the cleaning device 2 relative to the pipe PI can be determined, it becomes possible to change the direction of the cleaning device 2 so that it faces in a specific direction relative to the pipe PI, move the cleaning device 2 along the pipe PI, or move the cleaning device 2 perpendicular to the pipe PI.

The above configuration therefore makes it possible to control the movement of the cleaning device 2 based on images captured by the imaging device. Furthermore, the above configuration is more versatile than the technology of Patent Document 1 in that it does not require special components such as a stop position correction arm. Therefore, the above configuration has the effect of enabling versatile movement control of the cleaning device 2.

(Overview of pipe detection and tilt angle determination)
Fig. 3 is a diagram showing an overview of pipe detection and inclination angle determination by the control device 1. Fig. 2 shows a top view of the cleaning device 2 positioned on the pipe PI and an image IMG captured by the image capture device 271 provided in the cleaning device 2. Note that the appearance of the cleaning device 2 is simplified compared to Fig. 2. This also applies to Fig. 3 and subsequent figures.

The cleaning device 2 shown in Figure 3 has the front side of the cleaning device 2 tilted to the left with respect to the extension direction of the pipe PI. A camera 271 is also provided at the front of the cleaning device 2. This camera 271 is housed in the housing 27 in Figure 2 and is positioned so as to capture images of the lower side of the cleaning device 2. The camera 271 may be any device capable of capturing images in which the outline of the pipe PI can be recognized, and may be, for example, a depth camera.

In the image IMG captured by the image capture device 271 provided in the cleaning device 2 under these conditions, the pipe PI appears tilted upward to the right, as shown. As will be described in detail below, the pipe detection unit 101 detects a line segment L1 that forms the outer edge of the pipe PI from the image IMG. The angle identification unit 102 then identifies the inclination angle of the line segment L1 detected by the pipe detection unit 101 relative to a line segment L2 as the inclination angle of the pipe PI. Note that line segment L2 is a line segment that is perpendicular to the top and bottom sides of the image IMG.

When the front-to-rear direction of the cleaning device 2 is parallel to the extension direction of the pipe PI, that is, when the line segment forming the detected outer edge of the pipe is parallel to line segment L2, Δθ will be zero. Using the direction in which Δθ is zero as the reference, the inclination angle of the pipe can be expressed, for example, with a right-side inclination being positive and a left-side inclination being negative. In this case, as shown in Figure 3, when the front side of the cleaning device 2 is inclined to the left, Δθ will be a positive value, and when the front side of the cleaning device 2 is inclined to the right, Δθ will be a negative value.

(Example of pipe detection)
FIG. 4 is a diagram showing a specific example of pipe detection by the control device 1. IMG1 shown in FIG. 4 is an image captured by the image capture device 271. In this image IMG1, in areas A1 and A2, etc., the outer edge of the pipe, which should be straight, is distorted and curved. This distortion occurs when the lens of the image capture device 271 is a wide-angle lens. Note that proximity sensors are captured in the lower left and right corners of image IMG1.

Since distortion hinders pipe detection, in the example of Figure 4, prior to pipe detection, the pipe detection unit 101 performs distortion correction on image IMG1 to generate distortion-corrected image IMG2. In image IMG2, the outer edges of the pipes in areas A1' and A2' corresponding to areas A1 and A2 are straight. The distortion pattern that occurs depends on the imaging device 271, so for example, a check pattern can be photographed in advance using the imaging device 271, and correction parameters can be created that will eliminate distortion in the photographed check pattern. Distortion can be corrected using such correction parameters.

Image IMG3 shown in Figure 4 is also an image captured by the imaging device 271, but interference fringes appear in areas such as A3 of image IMG3. This type of noise can also interfere with pipe detection. For this reason, the pipe detection unit 101 may perform noise removal prior to pipe detection. Image IMG4 shown in Figure 4 is image IMG3 after noise removal processing has been performed. Image IMG4 is a clear image with no interference fringes.

Any noise removal method can be used. For example, noise may be removed using a bilateral filter. A bilateral filter removes noise while leaving the edge areas in the image clearly visible, making it suitable as a noise removal filter to be used in preprocessing for pipe detection.

From the image that has undergone the distortion correction and noise removal described above, the pipe detection unit 101 detects the line segments that make up the outer edge of the pipe. Various edge detection methods can be used to detect these line segments. For example, the pipe detection unit 101 may perform edge detection using the Canny method. In the Canny method, the target image is converted to grayscale, and then parts of the converted image where the change in brightness exceeds a threshold are detected as edges.

Figure 4 shows edge image IMG5, which is the result of edge detection using the Canny method from image IMG4. Edge image IMG5 is a binarized image in which various edges, including the outer edge of the pipe, are represented by white line segments, and the area other than the edges is a black background.

The pipe detection unit 101 then detects straight lines from the edge image IMG5. The detected straight lines include the edges of the outer edges of the pipes, so the process of detecting straight lines can be said to be a process of detecting pipes.

Any method can be used to detect straight lines from the edge image IMG5. For example, the pipe detection unit 101 may detect straight lines using a Hough transform. In this case, the pipe detection unit 101 expresses the equation of the straight line to be detected as ρ = x cos θ + y sin θ, and finds the (ρ, θ) pair, i.e., the polar coordinates, of the line on which a predetermined number or more edge pixels (white pixels in the edge image IMG5) are located.

As a result, each straight line (each straight line consisting of a predetermined number of white pixels or more) contained in the edge image IMG5 that is equal to or longer than a predetermined length is represented by polar coordinates (ρ, θ). Note that, in order to minimize the detection of straight lines other than the outer edge of the pipe to be detected, it is preferable to perform line detection while avoiding the image edges of the edge image IMG5 where anything other than the outer edge of the pipe is reflected. The area to be detected for straight lines can be determined in advance.

Next, the angle identification unit 102 identifies the tilt angle of the pipe from the straight line detected by the pipe detection unit 101. More specifically, the angle identification unit 102 identifies the tilt of the straight line expressed in polar coordinates (ρ, θ) in an x-y coordinate system, i.e., a Cartesian coordinate system. Specifically, the angle identification unit 102 finds two points on the straight line expressed in polar coordinates (ρ, θ) and finds the distance x1 in the x-axis direction and the distance y1 in the y-axis direction between those two points. Here, if the tilt to be identified is φ, then tanφ = y1/x1 holds, and therefore the angle identification unit 102 can identify the tilt φ from the found values of x1 and y1. However, 0 < φ < 180° must be satisfied.

The angle identification unit 102 performs the above process on all straight lines detected by the pipe detection unit 101 to identify the inclination of each detected straight line. Here, the angle identification unit 102 excludes any identified angles that exceed a threshold value. The angle identification unit 102 then identifies the average value of the remaining angles that were not excluded as the inclination angle of the pipe.

In addition, pipe detection and inclination angle determination may be performed multiple times in a time series, and the moving average of each determination result may be determined as the pipe inclination angle. In this case, for example, photography using the imaging device 271, and pipe detection and inclination angle determination from the captured images may be performed at a predetermined cycle (e.g., several Hz to several tens of Hz), and the moving average of the determination results over multiple cycles may be determined as the pipe inclination angle over those multiple cycles.

(Example of control: turning)
The movement control unit 104 can rotate the cleaning device 2 based on the tilt angle identified by the angle identification unit 102 as described above, and orient the cleaning device 2 in a desired direction. Specifically, the movement control unit 104 sets a target tilt angle _ΔθT , and then rotates the cleaning device 2 so that the tilt angle identified by the angle identification unit 102 becomes Δθ= _ΔθT . For example, when rotating the cleaning device 2 so that the front-to-rear direction is parallel to the extension direction of the pipe, the movement control unit 104 sets _ΔθT =0. Then, the movement control unit 104 rotates the cleaning device 2 until the tilt angle identified by the angle identification unit 102 becomes Δθ= _ΔθT =0.

However, there is a time lag between when the control device 1 sends a control signal to the cleaning device 2 and when the cleaning device 2 operates in accordance with that control signal. For this reason, it is preferable for the movement control unit 104 to perform rotation control that takes this time lag into account.

For example, the movement control unit 104 may perform turning control using the tilt angle Δθ _S expressed by the following formula as a threshold value. Note that _td is the dead time from the transmission of the control signal until the cleaning device 2 starts operating, γ is the angular velocity of the cleaning device 2 during turning, V is the traveling speed of the cleaning device 2 (the speed at which the cleaning device 2 travels at the rotational speed of the motor during turning), and a is the traveling acceleration of the cleaning device 2 (the acceleration when the cleaning device 2 travels at the rotational speed of the motor during turning). The coefficient of γ (½) in the following formula is set on the assumption that the angular velocity decreases linearly during the period from turning until the turning stops. The coefficient of γ is not limited to ½, as long as it corresponds to the pattern of fluctuations in the angular velocity during the period until the turning stops.
Δθ _S =t _d *γ+V/a*γ/2
Specifically, the movement control unit 104 outputs control signals for turning right, turning left, and stopping (ending turning) in accordance with the following conditions: Note that the sign of Δθ in the following conditional equation and the sign of Δθ _S on the right side change depending on how the reference axis is defined.
Δθ>Δθ _T -Δθ _S : Turn right Δθ<Δθ _T +Δθ _S : Turn left |Δθ|=Δθ _T +Δθ _S : Stop (end of turn)
The flow of processing during rotation is, for example, as follows: Note that the following describes rotation control when the cleaning device 2 is made parallel to the pipe, that is, rotation control when Δθ _T =0.

First, the movement control unit 104 calculates the angular velocity γ of the cleaning device 2, and then multiplies the angular velocity γ by a predetermined dead time _td to calculate the angle that changes during the time lag, i.e., the value of the first term on the right side of the above calculation formula for Δθ _S. The movement control unit 104 also calculates the value of the second term on the right side of the above calculation formula for Δθ _S from the travel speed V and acceleration a of the cleaning device 2, thereby obtaining the value of Δθ _S.

Next, the movement control unit 104 determines whether the magnitude of the tilt angle Δθ (|Δθ|) identified by the angle identification unit 102 is greater than Δθ _S. If |Δθ|≦Δθ _S , the movement control unit 104 outputs a signal to stop turning. If |Δθ|≦Δθ _S , this is because the required accuracy is met.

On the other hand, the movement control unit 104 rotates the cleaning device 2 if |Δθ|> _ΔθS . The conditions for determining the rotation direction are as described above. That is, the movement control unit 104 outputs a control signal for right rotation if Δθ> _ΔθS , and for left rotation if Δθ< _-ΔθS . The movement control unit 104 then returns to determining whether |Δθ|>ΔθS. By repeating this process until |Δθ|≦ _ΔθS , rotation that satisfies the required accuracy is achieved, taking into account the change in tilt angle during the dead time _{td ( td} *γ) and the change in tilt angle between sending the stop signal and stopping (V/a*γ/2).

(Outline of the method for calculating the amount of deviation)
Fig. 5 is a diagram showing an outline of a method for calculating the amount of deviation by the deviation amount calculation unit 103. Fig. 5 shows a plan view of the cleaning device 2 positioned above the pipes PIL and PIR, and a cross-sectional view taken along line A-A' in the plan view. Note that the cross-sectional view does not show anything other than the proximity sensors 272L and 272R provided on the cleaning device 2 and the pipes PIL and PIR.

As described above, the amount of deviation is calculated using the detection values of a pair of proximity sensors that detect pipe PI and are attached to symmetrical positions on the cleaning device 2. In the example in Figure 5, proximity sensors 272L and 272R are provided at the front of the cleaning device 2. Note that proximity sensor 272L is provided on the left side of center line L3, which bisects the cleaning device 2 in the left-right direction, and proximity sensor 272R is provided on the right side of center line L3. These sensors are housed in housing section 27 (see Figure 2).

As shown in the plan view, center line L3 is parallel to pipes PIL and PIR, but is shifted to the left by Δy from midpoint line L4, which indicates the center position between pipes PIL and PIR. The shift amount calculation unit 103 calculates this shift amount Δy based on the detection values of proximity sensors 272L and 272R.

More specifically, the deviation amount calculation unit 103 calculates Δy by utilizing the fact that when Δy is zero, the difference between the detection value of the proximity sensor 272L and the detection value of the proximity sensor 272R is zero, and that when Δy is other than zero, the difference is not zero.

For example, in the example of Figure 5, as shown in the cross-sectional view of the same figure, the distance from proximity sensor 272L to pipe PIL is shorter than the distance from proximity sensor 272R to pipe PIR. Therefore, the difference between the detection values of proximity sensor 272L and proximity sensor 272R will not be zero. For example, if proximity sensors 272L and 272R are used, whose detection values decrease as the distance to the target object decreases, the difference between the detection values will be a negative value.

As a result, because the difference in the detection values between proximity sensors 272L and 272R is a negative value, it can be seen that line L5, which indicates the center position between proximity sensors 272L and 272R, is shifted to the left of line L6, which indicates the center position between pipes PIL and PIR. Furthermore, because the magnitude of the difference in the detection values reflects the magnitude of the shift, Δy can be calculated from the magnitude of the difference in the detection values.

(Proximity sensor placement)
The pair of proximity sensors 272L and 272R are preferably arranged so that when one is located directly above a pipe, the other is at the detection limit of the pipe adjacent to that pipe. The reason for this arrangement will be explained with reference to Figures 6 and 7.

Figure 6 is a diagram illustrating the positional relationship between the proximity sensor 272 and the pipe PI, and the relationship between the detection value of the proximity sensor 272. Note that the proximity sensor 272 is the same sensor as the proximity sensors 272L and 272R.

Figure 6 shows the relationship between the amount of deviation (in mm) from the initial position of the proximity sensor 272 and the measurement value of the proximity sensor 272 (detection value of the proximity sensor 272, in volts). Note that the initial position is approximately 5 mm to the left of the position directly above the pipe PI.

As shown in the figure, when the proximity sensor 272 is moved to the right from its initial position, the detected value is approximately 1.2 V when the proximity sensor 272 is positioned directly above the pipe PI, i.e., when the proximity sensor 272 is closest to the pipe PI. This value is the minimum output value of the proximity sensor 272.

After this, as the proximity sensor 272 is moved further to the right, the detection value of the proximity sensor 272 increases, and when the tube PI is outside the detection range, the detection value (voltage value) is approximately 5.0V. This value is the maximum output value of the proximity sensor 272, and is the value that marks the boundary of the detection range of the proximity sensor 272. In other words, if the detection value (voltage value) is less than 5.0V, it can be said that the proximity sensor 272 is detecting the tube PI, and if it is 5.0V, it can be said that the proximity sensor 272 is not detecting the tube PI.

The relationship between the position of the proximity sensor 272 and the detection value has different patterns when the distance to the pipe PI is relatively short and when it is close to the detection range boundary. When the distance to the pipe PI is relatively short, the relationship between the position of the proximity sensor 272 and the detection value can be formulated as a function by approximation. This function is shown as an "approximation curve" in Figure 6.

In Figure 6, D represents the difference in deviation between the voltage value outside the detection range (approximately 5.0 V) on this approximation curve and the deviation when the proximity sensor 272 is closest to the tube PI (when the proximity sensor 272 is located directly above the tube PI). When the proximity sensor 272 moves a distance D to the right from a state where it is located directly above the tube PI, the detection value of the proximity sensor 272 does not reach the maximum value, but is close to it.

For this reason, if D is considered the detection limit distance, and the proximity sensor 272 is located in the area from a position directly above the pipe PI to a position D away to the right of that, the relationship between the detection value of the proximity sensor 272 and the amount of deviation can be approximated by an approximate curve such as that shown in Figure 6. Furthermore, when the pipe PI is located farther away than the detection limit of the proximity sensor 272, the detection value of the proximity sensor 272 can be approximated to a constant value (approximately 5.0 V).

In addition, experiments were conducted by changing the distance between the proximity sensor 272 and the pipe PI when the proximity sensor 272 was closest to the pipe PI. However, the shorter the distance, the smaller the detection value at closest proximity; the detection limit position did not change. From these experimental results, it can be said that it is reasonable to approximate the proximity portion with an approximation curve such as that shown in Figure 6, and that it is also reasonable to set the detection value at a position farther away than the detection limit position calculated from the function representing that approximation curve to a constant value (the maximum value of the proximity sensor 272).

7 is a diagram illustrating the relationship between the positional relationship between the proximity sensors 272L and 272R and the pipes PIL and PIR, and the approximate formula for the detection values of the proximity sensors 272L and 272R. In FIG. 7, the amount of deviation is set to zero when the left proximity sensor 272L is located directly above the left-hand pipe PIL of the two adjacent pipes. The figure also shows an approximate formula that represents the relationship between the amount of deviation x from the above position when the proximity sensors 272L and 272R are moved horizontally to the right from that position and the detection values _V1 and _V2 of the proximity sensors 272L and 272R.

As shown,
V ₁ =a(x-p) ² +q
V ₂ =a(xp-D) ² +q
where a, p, and q are constants determined by the distance between the proximity sensors 272, the diameter of the pipe PI, the pitch of the pipe PI, and the distance between the proximity sensor 272 and the pipe PI. Also, D is the difference between the amount of deviation when _V1 = _V1MAX (approximately 5.0 V) and the amount of deviation when _V1 = _V1MIN (approximately 1.2 V) in the approximate curve of _V1 .

The proximity sensors 272L and 272R are positioned using the distance D thus determined. Specifically, when the proximity sensor 272L is located directly above the pipe PIL, the horizontal distance from the proximity sensor 272R to the position directly above the pipe PIR is set to D. Furthermore, when the proximity sensor 272R is located directly above the pipe PIR, the horizontal distance from the proximity sensor 272L to the position directly above the pipe PIL is also set to D. To achieve this positional relationship, the distance d between the proximity sensors 272L and 272R can be set to d = (P - D), where P is the distance between the pipes PIL and PIR.

With this arrangement, when proximity sensor 272L is located directly above tube PIL, detected value _V1 is the minimum value _V1MIN (approximately 1.2 V). At this time, proximity sensor 272R is located at the detection limit of tube PIR, so _V2 is the maximum value _V2MAX (approximately 5.0 V). On the other hand, when proximity sensor 272R is located directly above tube PIR, detected value _V2 is the minimum value _V2MIN (approximately 1.2 V). At this time, proximity sensor 272L is located at the detection limit of tube PIL, so _V1 is the maximum value _V1MAX (approximately 5.0 V).

Furthermore, in the section from when the proximity sensor 272L is located directly above the pipe PIL to when the proximity sensor 272R is located directly above the pipe PIR, the proximity sensor 272L detects the pipe PIL and the proximity sensor 272R detects the pipe PIR. The difference in the detection values of the proximity sensors 272L and 272R in this section can be expressed as follows:
V ₁ -V ₂ ={a(x-p) ² +q}-{a(x-p-D) ² +q}
=2aDx-a(2pD+D ² ) ²
5, when the detection values of proximity sensors 272L and 272R are equal, that is, when V ₁ - V ₂ = 0, the amount of deviation also becomes zero. Therefore, the function representing the amount of deviation can be expressed as the following linear function by setting the second term on the right side of the above equation to zero.

x=(V ₁ -V ₂ )/2aD
Now, let's assume that proximity sensor 272R is positioned directly above tube PIR, and then proximity sensors 272L and 272R move further to the right. In this state, proximity sensor 272L is outside the detection range of tube PIR, so V ₁ is a constant of V _1MAX (approximately 5.0 V). On the other hand, proximity sensor 272R is within the detection range of tube PIR, and is expressed as V ₂ = a(x-p-D) ² + q. Therefore, the difference in the detection values of proximity sensors 272L and 272R in this state can be expressed as follows:
V ₁ −V ₂ =V _1MAX −a(xp−D) ² −q
As described above, the deviation is set to zero when V ₁ - _{V 2} = 0. In other words, -a(2pD+D ² ) ² = 0. Since a≠0 and D≠0, 2p+D=0. Therefore, when proximity sensor 272L is outside the detection range of tube PIL and proximity sensor 272R is within the detection range of tube PIR, the function representing the deviation is expressed as follows: x={-(V ₁ - V ₂ + q-V _1MAX )/a} ^1/2 +D/2
As described above, the difference between the detection values of proximity sensors 272L and 272R can be expressed by simple approximations for the cases where only one of proximity sensors 272L and 272R detects a pipe and where both detect a pipe. By using these approximations, a function for determining the amount of deviation can be derived from the difference between the detection values of proximity sensors 272L and 272R.

Fig. 8 is a diagram showing a function for calculating the amount of deviation from the difference (V ₁ -V ₂ ) in the detection values of proximity sensors 272L and 272R. The vertical axis of the graph in Fig. 8 represents the amount of deviation (unit: mm) of proximity sensor 272L from the position directly above pipe PIL, and the horizontal axis represents the difference (voltage difference, unit: V) in the detection values of proximity sensors 272L and 272R.

The function shown in Fig. 8 is a straight line (linear function) in the section where the amount of deviation is from 0 to approximately 15 mm. Specifically, the function in this section is a linear function in which the amount of deviation increases in proportion to the voltage difference (V ₁ - _{V 2} ), as shown in equation (1) in Fig. 8.

The function shown in Fig. 8 is a curve in the section where the deviation amount is from about 15 to 45 mm. Specifically, the function in this section is a function in which the deviation amount decreases according to the value of (V ₁ - V ₂ ) ^1/2 , as shown in formula (2) in Fig. 8. Note that in Fig. 8, V _1MAX = 5. In other words, the "5" in (V ₁ - _{V 2} + q - 5) in formula (2) is V _1MAX .

That is, when both proximity sensors 272L and 272R detect a pipe (when _V1 ≠ _V1MAX and _V2 ≠ _V2MAX ), deviation calculation unit 103 calculates deviation x using the above formula (1). Also, when only one of proximity sensors 272L and 272R detects a pipe, deviation calculation unit 103 calculates deviation x using the above formula (2). Here, when only one of proximity sensors 272L and 272R detects a pipe, this refers to ( _V1 = _V1MAX and _V2 ≠ _V2MAX ) or ( _V2 = _V2MAX and _V1 ≠ _V1MAX ).

In this way, by positioning proximity sensor 272R so that it is at the detection limit of pipe PIR when proximity sensor 272L is positioned directly above pipe PIL, and proximity sensor 272L is at the detection limit of pipe PIL when proximity sensor 272R is positioned directly above pipe PIR, the amount of deviation can be calculated using the simple functions of equations (1) and (2).

Depending on the distance between the proximity sensors 272L and 272R and the pipes PIL and PIR, the value in the root of equation (2) may be negative. In this case, the deviation amount calculation unit 103 sets the first term on the right side of equation (2) to zero. Furthermore, depending on whether the value of ( _V1 - _V2 ) is positive or negative, in other words, the magnitude relationship between _V1 and _V2 , it can also be determined whether the center position of the adjacent pipes and the center position of the proximity sensors 272L and 272R are misaligned to the left or right. Specifically, if _V1 > _V2, it can be determined that there is a rightward deviation, and if _V2 > _V1, it can be determined that there is a leftward deviation.

As described above, the control device 1 includes a deviation amount calculation unit 103 that calculates the amount of deviation of the cleaning device 2 from a predetermined reference position based on the detection values of a pair of proximity sensors 272L and 272R that detect pipes and are attached to the cleaning device 2. The movement control unit 104 then controls the movement of the cleaning device 2 based on the tilt angle identified by the angle identification unit 102 and the amount of deviation calculated by the deviation amount calculation unit 103.

More specifically, a pair of proximity sensors 272L and 272R are positioned so that when one is positioned directly above a pipe, the other is at the detection limit of the pipe adjacent to that pipe. The deviation amount calculation unit 103 then calculates the deviation amount using formula (2), which is an approximation of the relationship between the difference in detection values between the pair of proximity sensors 272L and 272R and the deviation amount.

As described above, if proximity sensors 272L and 272R are attached to cleaning device 2 at symmetrical positions on the left and right sides to detect two parallel-arranged pipes, when cleaning device 2 is located in the center of the two pipes, the distance from proximity sensor 272L to pipe PIL will be equal to the distance from proximity sensor 272R to pipe PIR. In this case, the output values of proximity sensors 272L and 272R will be the same or nearly the same.

On the other hand, when the cleaning device 2 is positioned away from the center position of the two pipes, the output values of the proximity sensors 272L and 272R will be different. In this case, the difference in the output values of the proximity sensors 272L and 272R will be a value that corresponds to the magnitude of the deviation between the center position of the cleaning device 2 in the left-right direction and the center position between the pipes PIL and PIR.

Therefore, the amount of deviation of the cleaning device 2 from its reference position can be calculated based on the detection values of the proximity sensors 272L and 272R attached at symmetrical positions on the left and right of the cleaning device 2. For example, it is also possible to calculate the amount of deviation between the center position of the cleaning device 2 in the left-right direction and the center position between multiple tubes arranged in parallel.

Then, by controlling movement using this amount of deviation, the cleaning device 2 can be aligned to a predetermined position relative to the pipe. For example, by controlling the movement of the cleaning device 2 so that the amount of deviation is zero, the cleaning device 2 can be aligned to the center position between the pipes.

As shown in Figure 6, the output characteristics of the proximity sensor 272 differ when the detection target is near the proximity sensor 272 and when the detection target is located far from the proximity sensor 272, i.e., near the detection limit. Therefore, when approximating the output characteristics when the detection target is near the proximity sensor 272 with a quadratic equation, strictly speaking, the output characteristics when the detection target is near the detection limit must be approximated with a different equation.

However, in this case, the output characteristics of proximity sensor 272 are divided into three parts: a quadratic equation, the other equation mentioned above, and a constant. If two proximity sensors, proximity sensors 272L and 272R, are used, the equation expressing the difference between their detected values becomes complicated, and the need for multiple case distinctions increases, making the calculations more complicated.

The proximity sensors 272L and 272R of the cleaning device 2 are positioned so that when one is directly above a pipe, the other is at the detection limit of the pipe adjacent to it. The detection value at the detection limit is then approximated as the maximum detection value. This allows the amount of deviation to be calculated by a simple calculation: when both proximity sensors 272L and 272R are detecting the pipe, the above formula (1) is used; when only one of proximity sensors 272L and 272R is detecting the pipe, the above formula (2) is used.

In this embodiment, an example is described in which proximity sensors 272L and 272R are attached at symmetrical positions on the left and right sides of the cleaning device 2, but this is not limiting. Proximity sensors 272L and 272R may be positioned in a predetermined direction and at a predetermined distance from a reference position (e.g., the center position) of the cleaning device 2. For example, one proximity sensor may be positioned a predetermined distance forward of the center position of the cleaning device 2, and the other proximity sensor may be positioned a predetermined distance behind it.

Furthermore, the detector for detecting the pipe is not limited to the proximity sensor 272. For example, any detector that can detect the pipe without contact, such as a distance sensor (e.g., a laser distance sensor) or an ultrasonic sensor, can be applied. Furthermore, the number of detectors installed is not limited to two, but may be three or more, and multiple types of detectors may be used in combination.

Furthermore, the approximation formula for the relationship between the difference in detection values of the proximity sensors and the amount of deviation is not limited to the examples of formulas (1) and (2). For example, depending on the type and placement of the detector used to detect the pipe, approximation can also be performed using higher-order functions (including quadratic and cubic functions), exponential functions, logarithmic functions, or combinations thereof. However, the configuration using formulas (1) and (2) described above is preferable because it has the advantage of allowing the amount of deviation to be calculated using simple calculations.

(Example of control: centering)
As an example of control of the cleaning device 2 based on the above-mentioned amount of misalignment and tilt angle, we will explain centering control of the cleaning device 2. Centering refers to moving the cleaning device 2 to a central position between the pipes, and by centering, the pantograph 25 can be lowered between the pipes to perform cleaning (see FIG. 2).

Figure 9 is a diagram showing an example of the operation of the cleaning device 2 during centering. Figure 9 shows a view from above of the cleaning devices 2 on pipes PI1 to PI3, which are arranged parallel to each other and at equal intervals. Also in Figure 9, the center position between pipes PI1 and PI2 is indicated by dashed line L7. Centering is complete when dashed line L8, which divides the cleaning device 2 in half horizontally, coincides with dashed line L7.

To center the cleaning device 2, the movement control unit 104 first places the cleaning device 2 parallel to the pipes PI1 to PI3. Specifically, the movement control unit 104 acquires the latest tilt angle Δθ identified by the angle identification unit 102, sets the target tilt angle _ΔθT to zero, calculates the threshold value _ΔθS , and determines whether |Δθ|= _ΔθS (ST1). In the example of FIG. 9, |Δθ|> _ΔθS . In this case, the movement control unit 104 rotates the cleaning device 2 to the left until |Δθ|= _ΔθS (ST2).

When |Δθ|= _ΔθS , the movement control unit 104 acquires the latest deviation Δy calculated by the deviation amount calculation unit 103 and determines whether |Δy|= _yS (ST3). _yS may be set to, for example, a value obtained by adding a predetermined margin to 0. The margin is determined according to the pipe pitch, pipe diameter, pantograph width, and the like. The margin is considered to be approximately several mm in most cases. In the example of FIG. 9, |Δy|> _ΔyS . In this case, the movement control unit 104 performs control to bring Δy closer to zero.

Specifically, the movement control unit 104 moves the cleaning device 2 forward or backward while orienting it in a direction inclined relative to the extension direction of the pipes PI1 to PI3, thereby bringing Δy closer to zero. In this case, it is anticipated that centering by moving forward or backward only may result in the device moving away from the starting position for straight-ahead travel. For this reason, it is preferable to perform centering using a multi-stage process that alternates between moving forward and backward.

When alternately repeating forward and backward movement, the movement control unit 104 determines whether the previous movement was forward or backward movement. It also determines whether to move leftward or rightward based on the value of Δy. For example, if the Δy value is output as positive at positions to the right of Δy = 0 and as negative at positions to the left of Δy = 0, the movement control unit 104 can determine to move leftward if the Δy value is positive, or to move rightward if the Δy value is negative.

The movement control unit 104 then determines the turning direction based on these determination results. Specifically, when moving left and the previous movement was forward, the movement control unit 104 determines to make a right turn. Furthermore, when moving left and the previous movement was reverse, the movement control unit 104 determines to make a left turn. The same is true when moving right; the movement control unit 104 determines to make a left turn if the previous movement was forward, and to make a right turn if the previous movement was reverse. In the example of Figure 9, it is assumed that it is necessary to move right and the previous movement was forward, so the movement control unit 104 determines to make a left turn (ST4).

After determining the turning direction, the movement control unit 104 transmits a turning instruction in the determined direction to the cleaning device 2. The target turning angle _ΔθT may be determined in advance. That is, after transmitting the turning instruction, the movement control unit 104 instructs the cleaning device 2 to end turning at the timing when the angle specified by the angle specifying unit 102 becomes Δθ= _ΔθT (ST5).

After the above-mentioned turning is completed, the movement control unit 104 moves the cleaning device 2 forward or backward. Specifically, if the previous movement was a forward movement, a backward movement command is sent to the cleaning device 2, and if it was a backward movement, a forward movement command is sent to the cleaning device 2 (ST6). The method for determining the forward or backward movement time at this time, i.e., the running time of the cleaning device 2, will be described later with reference to Figure 10.

(How to calculate driving time)
Fig. 10 is a diagram illustrating a method for calculating the travel time of the cleaning device 2 during centering. Fig. 10 shows the cleaning device 2 as viewed from above. Note that the cleaning device 2 is depicted small. In Fig. 10, the center position of the parallel pipes PIL and PIR is indicated by a dashed line L9, and the line dividing the cleaning device 2 in half in the left-right direction is indicated by a dashed line L10.

As shown in the figure, the angle between dashed lines L9 and L10 is _ΔθT , the amount of deviation of the cleaning device 2 from dashed line L9 is Δy, and the movement distance from the center position of the cleaning device 2 to a position on dashed line L9 when moving forward along dashed line L10 is _LP .

At this time, the movement control unit 104 may calculate the travel time of the cleaning device 2 using formula (3) shown in FIG. 10 . Note that V in formula (3) is the travel speed of the cleaning device 2. Also, in formula (3), Δy/sin(Δθ _T ) = L _P. In other words, formula (3) is an equation that calculates the travel time by using L _P * (1 + k _b ) = Δy/sin(Δθ _T ) * (1 + k _b ), which is L _P multiplied by (1 + k _b ), rather than L _P as the travel distance of the cleaning device 2. Because pipes are generally slippery, by calculating the travel time after adding the travel distance in this way, it is possible to make the cleaning device travel a distance closer to L _P.

The above _kb is a bias value. As shown in Fig. 10, kb may be expressed as _kb = ( _kp + _kn )/2, that is, the arithmetic mean value of the bias value _kp during the previous movement and the bias value _kn during the current movement. Note that, as shown in Fig. 10, the current bias value _kn is the ratio ( _Δyn / _Δyp ) of the positional deviation amount _Δyn after the most recent movement to the _positional deviation amount _Δyp before the movement.

For example, EX1 in Figure 10 shows an example of centering by moving three stages from a state where the amount of positional deviation is Δy1. In EX1, the center position of the cleaning device 2 when the amount of positional deviation is Δy1 is indicated by point P1. Furthermore, the amount of positional deviation after the first stage of movement is Δy2, and the center position of the cleaning device 2 at this time is indicated by point P2. The amount of positional deviation after the second stage of movement is Δy3, and the center position of the cleaning device 2 at this time is indicated by point P3.

In the first stage of movement, the movement control unit 104 calculates the travel time using equation (3) with a default bias value _kd . Here, Δy1 can be substituted for Δy in equation (3). For _kd , for example, the last used bias value _kb can be applied, or a predetermined value can be applied. Note that a predetermined constant smaller than Δy1 can be substituted for Δy to ensure that centering is performed in multiple stages. This constant indicates the upper limit of the shift width in one stage of movement, and can be set in advance depending on the space on the pipe, etc. The travel time is calculated using this constant until the amount of shift calculated by the shift amount calculation unit 103 becomes equal to or less than this constant.

In the second-stage movement, the movement control unit 104 calculates the ratio (Δy2/Δy1) of the positional deviation amount Δy2 after the first-stage movement to the positional deviation amount Δy1 before the first-stage movement as the current bias value _kn . Then, the movement control unit 104 sets the arithmetic mean value of the calculated _kn and the previously applied bias value _kd as the bias value for the second-stage movement. In this way, in the second-stage movement, control of the movement distance according to the deviation amount before and after the first-stage movement is realized.

For the third-stage movement, the movement control unit 104 calculates the ratio (Δy3/Δy2) of the positional deviation amount Δy3 after the second-stage movement to the positional deviation amount Δy2 before the second-stage movement as the current bias value _kn . Then, the movement control unit 104 sets the arithmetic mean value of the calculated _kn and the previously applied bias value _kb as the bias value for the third-stage movement. As a result, for the third-stage movement, control of the movement distance according to the deviation amounts before and after the second-stage movement is realized.

In EX1, the dashed line L9, i.e., the center position between the pipes PIL and PIR, is reached after three stages of movement, but if the dashed line L9 is not reached even after the third stage of movement, the movement control unit 104 similarly performs a fourth stage of movement.

In this way, the movement control unit 104 may perform centering while repeatedly moving forward and backward and updating the bias value. Note that depending on the condition of the pipe, it is possible that the movement may pass over the dashed line L9. In such a case, the sign of the deviation amount Δy is reversed before and after the movement, and k _n = ( _Δyn / _Δyp ) becomes a negative value. In this case, the movement control unit 104 may perform movement control with k _n = 0.

As described above, when moving the cleaning device 2 a predetermined distance, the movement control unit 104 may move the predetermined distance in multiple stages. In this case, it is preferable that the movement control unit 104 adjusts the movement distance in a subsequent stage of movement depending on the amount of deviation before and after the movement in the previous stage.

When cleaning device 2 is moved along a pipe, even if it is controlled to move forward at the same set speed for the same amount of time, variations in the distance traveled may occur due to factors such as the slipperiness of the pipe surface. Therefore, with the above configuration, when cleaning device 2 is moved a predetermined distance, the predetermined distance is moved in multiple stages, and the distance traveled in subsequent stages is adjusted depending on the deviation before and after the movement in the previous stage. This makes it possible to stably move cleaning device 2 a predetermined distance regardless of the surface condition of the pipe along which cleaning device 2 is moving.

10, adjustment is performed according to the amount of deviation before and after the immediately preceding movement, but adjustment may also be performed taking into account the amount of deviation before that. For example, in the third stage of movement in EX1, the arithmetic mean value of the bias value used in the first stage adjustment, the bias value used in the second stage adjustment, and the newly calculated b _n may be used as the bias value b _b . Also, a weighted mean value or the like may be used instead of the arithmetic mean value.

(Example of control: driving straight ahead)
After centering and cleaning at the center position between the pipes, the movement control unit 104 moves the cleaning device 2 in a straight line while maintaining the center position between the pipes, thereby enabling cleaning at another position on the same pipe.

To control the cleaning device 2 to travel in a straight line while maintaining the center position between the pipes, the tilt angle identified by the angle identification unit 102 can be used, for example. This is because if the tilt angle can be maintained at zero, the center position between the pipes can also be maintained.

Furthermore, if the deviation between the cleaning device 2 and the central position before straight-line travel is zero or a value close to zero, it is not difficult to keep the final deviation within the allowable range. On the other hand, if the deviation between the cleaning device 2 and the central position before straight-line travel is within the allowable range but is somewhat large, it becomes more difficult to keep the final deviation within the allowable range.

For this reason, before performing straight-line travel, the movement control unit 104 may determine whether the amount of deviation calculated by the deviation amount calculation unit 103 is equal to or less than a threshold value, and may perform different control when the amount of deviation is equal to or less than the threshold value and when it exceeds the threshold value. For example, in the former case, the movement control unit 104 may apply a high-speed straight-line mode in which the cleaning device 2 moves at high speed, and in the latter case, a low-speed straight-line mode in which the movement speed is slower than the high-speed straight-line mode.

(High-speed straight-line mode)
When the high-speed straight-line mode is applied, the movement control unit 104 may calculate the travel time as L/V, which is the distance to be traveled L divided by the travel speed V. After the cleaning device 2 starts traveling, the movement control unit 104 determines whether the tilt angle identified by the angle identification unit 102 exceeds the allowable value. If the movement control unit 104 determines that the tilt angle exceeds the allowable value, it adjusts the traveling direction of the cleaning device 2.

For example, if the cleaning device 2 is equipped with crawlers 22 as shown in Figure 2, the movement control unit 104 can adjust the direction of travel of the cleaning device 2 by increasing or decreasing the speed of one of the pair of left and right crawlers 22.

Now, let's say that the tilt angle Δθ identified by the angle identification unit 102 after starting straight-ahead travel exceeds the allowable value. In this case, if the tilt angle is positive (i.e., the cleaning device 2 is tilted to the left as in the example of Figure 3), the movement control unit 104 may multiply the speed of the left crawler 22 by α (α = 1 + |Δθ|). Also, if the tilt angle is negative (i.e., the cleaning device 2 is tilted to the right as opposed to the example of Figure 3), the movement control unit 104 may multiply the speed of the right crawler 22 by α.

When moving the cleaning device 2 backward, the opposite control is used, and the movement control unit 104 multiplies the speed of the right crawler 22 by α if the tilt angle is positive, and multiplies the speed of the left crawler 22 by α if the tilt angle is negative. By repeating this process until the movement of the distance L is complete, it becomes possible to move at a relatively high speed while maintaining a center position between the pipes.

(Low speed straight-line mode)
When the low-speed straight-line mode is applied, the movement control unit 104, as in the high-speed straight-line mode, sets the travel time to L/V, and checks the tilt angle until travel of the distance L is completed, and if the tilt angle exceeds the allowable value, performs control to increase the speed of one of the crawlers 22 by a factor of α. However, when the low-speed straight-line mode is applied, it is preferable that the movement control unit 104 also checks the amount of deviation, and adjusts the traveling direction of the cleaning device 2 if the amount of deviation exceeds the allowable value.

For example, the movement control unit 104 may perform the above control when the tilt angle identified by the angle identification unit 102 exceeds the allowable value, or may check the amount of deviation calculated by the deviation amount calculation unit 103 when the tilt angle identified by the angle identification unit 102 is equal to or less than the allowable value. Then, when the amount of deviation exceeds the allowable value, the movement control unit 104 may perform an operation to move the object toward the center position. Note that when the amount of deviation is equal to or less than the allowable value, the movement control unit 104 may simply return to the tilt angle confirmation process.

During the approaching operation, if the amount of deviation is a value indicating that the cleaning device 2 is shifted to the left of the central position, the movement control unit 104 may increase the speed of the left crawler 22 by β (β > α). The movement control unit 104 may then run the cleaning device 2 for a predetermined time with the speed of the left crawler 22 increased by β, and then increase the speed of the right crawler 22 by β, thereby returning the vehicle body angle. Furthermore, if the amount of deviation is a value indicating that the cleaning device 2 is shifted to the right of the central position, the movement control unit 104 may increase the speed of the right crawler 22 by β. The movement control unit 104 may then run the cleaning device 2 for a predetermined time with the speed of the right crawler 22 increased by β, and then increase the speed of the left crawler 22 by β, thereby returning the vehicle body angle. When moving the cleaning device 2 backward, the control is reversed; if the device is to the left, the movement control unit 104 causes the device 2 to travel for a predetermined time with the speed of the left crawler 22 multiplied by β, and if the device is to the right, the movement control unit 104 causes the device 2 to travel for a predetermined time with the speed of the right crawler 22 multiplied by β.

In this way, by running the left and right crawlers 22 at different speeds for a predetermined period of time, the amount of deviation of the cleaning device 2 can be reduced or eliminated. After running the cleaning device 2 for a predetermined period of time with the left and right crawlers 22 at different speeds, the movement control unit 104 can return the speeds of the left and right crawlers 22 to the same speed and return to the tilt angle confirmation process.

(Example of control: column movement)
When cleaning of the entire two adjacent pipes is completed, the movement control unit 104 causes the cleaning device 2 to move in a row. Note that the row in the row movement refers to the section between a certain pipe and a pipe adjacent to that pipe.

In the row movement, the movement control unit 104 first rotates the cleaning device 2 to the target angle Δθ _T based on the tilt angle identified by the angle identification unit 102, and then moves the cleaning device 2 forward after the rotation is complete. The travel time for this forward travel may be calculated, for example, using the following equation (4). Note that P is the distance between adjacent pipes, and V is the travel speed of the cleaning device 2.

(Travel time)=P*sin( _ΔθT )/V…(4)
Next, the movement control unit 104 returns the tilt angle of the cleaning device 2 to zero based on the tilt angle identified by the angle identification unit 102. Here, assume that the cleaning device 2 was positioned at the center of the row before moving and that the cleaning device 2 had moved forward by P*sin( _ΔθT ) due to the above-mentioned travel control. In this case, the amount of deviation of the cleaning device 2 from the center position of the row after movement at the time when the tilt angle is returned to zero is P*{cos( _ΔθT )} ² .

Next, the movement control unit 104 rotates the cleaning device 2 by (90-Δθ _T ) based on the tilt angle identified by the angle identification unit 102. In this way, the movement control unit 104 may rotate the cleaning device 2 so that the first and second rotation angles differ by 90°. Then, after the above rotation is completed, the movement control unit 104 moves the cleaning device 2 backward.

Here, if the cleaning device 2 was positioned in the center of the row before moving in line and moved forward by P*sin( _ΔθT ) during the first forward movement, then if the cleaning device 2 is moved backward by P*cos( _ΔθT ), the position of the cleaning device 2 after moving backward will be a position that is moved directly to the side by P from the position of the cleaning device 2 before moving in line. Therefore, the traveling time for backward movement is calculated using the following formula (5).

(Traveling time)=P*cos( _ΔθT )/V…(5)
The row movement is achieved by a combination of turning, moving forward, turning, and moving backward as described above. Note that the cleaning device 2 may be moved backward after the first turning, in which case the cleaning device 2 may be moved forward after the second turning. The turning direction may be determined depending on which row the cleaning device 2 is to be moved to and the direction of travel of the cleaning device 2. For example, when moving to the left row and moving the cleaning device 2 forward after the first turning, the cleaning device 2 may be turned left.

Note that even if the cleaning device 2 is moved forward or backward for the above-mentioned travel time, it is possible that the cleaning device 2 will not move to the desired position due to factors such as the slipperiness of the pipe surface. For this reason, the deviation amount calculation unit 103 may calculate the deviation amount Δy from the center position of the line after the first travel. The movement control unit 104 may then adjust the second travel time according to the deviation amount Δy calculated by the deviation amount calculation unit 103. In other words, the movement control unit 104 may calculate the second travel time using the deviation amount Δy calculated after the first travel has ended.

Furthermore, once the above processing is completed, the deviation amount calculation unit 103 calculates the deviation amount Δy from the center position of the column after movement. If the deviation amount Δy exceeds the allowable range, the above-mentioned centering is performed.

(Processing before starting control of cleaning device)
For example, if the photographing device 271 is not photographing normally, the angle identification unit 102 cannot identify an appropriate tilt angle, and the movement control unit 104 cannot perform appropriate control.

Therefore, prior to performing the various controls described above, the movement control unit 104 may acquire the tilt angle identified by the angle identification unit 102 and determine whether the tilt angle is within a preset invalid range. If the tilt angle acquired in chronological order from the angle identification unit 102 is within the invalid range a predetermined number of times in succession, the movement control unit 104 may send an abnormality signal and terminate control of the cleaning device 2. This makes it possible to prevent unintended control from being performed.

(Regarding determination of the action to be performed by the cleaning device)
As described above, the control device 1 can perform operations such as centering, moving in a line, and traveling straight ahead. The determination of which of these operations to perform may be made by the control device 1 itself, or by another control device provided upstream of the control device 1. In the latter case, the upstream control device determines the operation to be performed by the cleaning device 2, for example, in accordance with an operator's operation, and notifies the control device 1 of the determined operation. The control device 1 operates the cleaning device 2 in accordance with this notification, and upon completion of the operation, notifies the upstream control device of a completion notification. After receiving the completion notification, the upstream control device determines the next operation to be performed and notifies the control device 1 of the determined operation. By repeating this process, the cleaning device 2 can clean the pipe. Controls that do not use the amount of displacement or tilt angle, such as control of the extension and retraction of the pantograph 25 and control of water discharge, may be performed by the control device 1 or the upstream control device.

(Process flow: Control based on tilt angle)
Of the processes executed by the control device 1, control based on the tilt angle identified by the angle identification unit 102 will be described with reference to Fig. 11. Fig. 11 is a flowchart showing an example of a method for controlling the cleaning device 2 based on the tilt angle. Note that it is assumed here that the cleaning device 2 is placed on a plurality of water pipes arranged in parallel, and that the water pipes are photographed by the photographing device 271 attached to the cleaning device 2.

In S11, the pipe detection unit 101 acquires an image captured by the imaging device 271 attached to the cleaning device 2. Then, in S12, the pipe detection unit 101 detects water pipes from the image acquired in S11. The method for detecting water pipes has already been explained, so the explanation will not be repeated here.

In S13, the angle identification unit 102 identifies the tilt angle of the pipe detected in S12. The method for identifying the tilt angle of the pipe detected in the image has already been described, so the description will not be repeated here.

In S14, the movement control unit 104 determines whether or not to control the cleaning device 2 based on the tilt angle identified in S13. If it is determined that control should not be performed (NO in S14), the processing in FIG. 11 ends. On the other hand, if it is determined that control should be performed (YES in S14), the processing proceeds to S15, where the movement control unit 104 controls the cleaning device 2 based on the tilt angle identified in S13, and the processing in FIG. 11 ends.

Note that various control contents can be applied to S15, and the determination criterion of S14 may be set appropriately depending on the control contents of S 15. For example, when performing rotation control to set the tilt angle of the cleaning device 2 to _ΔθT in S15, the movement control unit 104 determines in S14 that control is not necessary if the tilt angle specified in S13 is within an allowable range from _ΔθT , and determines that control is necessary if the tilt angle is not within the allowable range from _ΔθT .

For example, when the cleaning device 2 is traveling in a straight line, the movement control unit 104 determines in S14 that control is not necessary if the tilt angle identified in S13 is within the tolerance, and that control is necessary if it exceeds the tolerance. Then, in S15, the movement control unit 104 adjusts the speed of the crawler 22 so that the tilt angle identified in S13 becomes zero or approaches zero.

As described above, the control method for the cleaning device 2 executed by the control device 1 includes a pipe detection step (S12) for detecting pipes from images captured by the imaging device 271 attached to the cleaning device 2 that cleans the surfaces of pipes, an angle identification step (S13) for identifying the inclination angle of the pipe detected in the pipe detection step, and a movement control step (S15) for controlling the movement of the cleaning device 2 based on this inclination angle. This control method enables general-purpose movement control of the cleaning device 2 based on images captured by the imaging device 271.

(Process flow: Control based on deviation amount)
Of the processes executed by the control device 1, control based on the amount of deviation identified by the deviation amount calculation unit 103 will be described with reference to Fig. 12. Fig. 12 is a flowchart showing an example of a method for controlling the cleaning device 2 based on the amount of deviation. Note that it is assumed here that the cleaning device 2 is located between a pair of adjacent water tubes among a plurality of water tubes arranged in parallel.

In S21, the deviation amount calculation unit 103 acquires the detection values of a pair of proximity sensors 272L and 272R attached to symmetrical positions on the cleaning device 2. Then, in S22, the deviation amount calculation unit 103 calculates the deviation amount of the cleaning device 2 from a predetermined reference position based on the detection values acquired in S21. For example, the deviation amount calculation unit 103 calculates the deviation amount between the center position of the cleaning device 2 in the left-right direction and the center position between multiple tubes arranged in parallel.

As described above, when calculating this deviation amount, the approximation formula used by deviation amount calculation unit 103 to calculate the deviation amount differs depending on whether both proximity sensors 272L and 272R are detecting the water pipe (when both detection values are less than the maximum value) or whether only one is detecting the water pipe (when one detection value is the maximum value).

Specifically, when the detection values of both proximity sensors 272L and 272R are less than the maximum value, the deviation amount calculation unit 103 calculates the deviation amount by substituting the difference between those detection values into the above-mentioned formula (1). On the other hand, when the detection value of one of proximity sensors 272L and 272R is the maximum value and the detection value of the other is less than the maximum value, the deviation amount calculation unit 103 calculates the deviation amount by substituting the difference between those detection values into the above-mentioned formula (2).

In S23, the movement control unit 104 determines whether or not to control the cleaning device 2 based on the amount of deviation calculated in S22. If it is determined that control should not be performed (NO in S23), the processing in FIG. 12 ends. On the other hand, if it is determined that control should be performed (YES in S23), the processing proceeds to S24, where the movement control unit 104 controls the cleaning device 2 based on the amount of deviation calculated in S22, and the processing in FIG. 12 ends.

Various control operations can be applied to S24, and the criteria for S23 can be set appropriately depending on the control operation in S24. For example, when performing centering, the movement control unit 104 determines in S23 that control is not required if the amount of deviation calculated in S22 is within an allowable value, and that control is required if it exceeds the allowable value. Then, in S24, the movement control unit 104 performs control to reduce the amount of deviation. For example, the movement control unit 104 controls the cleaning device 2 to rotate to a predetermined tilt angle and then move forward or backward, or controls the speed of one of the left and right crawlers 22 to increase or decrease.

As described above, the control method for the cleaning device 2 executed by the control device 1 includes a deviation calculation step (S22) for calculating the deviation of the cleaning device 2 from a predetermined reference position based on the detection values of a pair of proximity sensors 272L and 272R attached at symmetrical positions on the cleaning device 2, which cleans multiple parallel-arranged pipes, and a movement control step (S24) for controlling the movement of the cleaning device 2 based on this deviation. Furthermore, the pair of proximity sensors 272L and 272R are positioned so that when one is positioned directly above a pipe, the other is at the detection limit of the pipe adjacent to that pipe. The deviation calculation step calculates the deviation using Equation (2), which is an approximation of the relationship between the difference in detection values of the pair of proximity sensors 272L and 272R and the deviation, or Equation (1), which is an approximation of the relationship between the difference in detection values of the proximity sensors 272L and 272R and the deviation using a linear function. This allows for versatile movement control of the cleaning device 2. In addition, the amount of deviation can be calculated through simple calculations using a simple approximation formula.

[Embodiment 2]
Other embodiments of the present invention will be described below. For ease of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

In this embodiment, an example will be described with reference to Figure 13 in which the deviation amount calculation unit 103 calculates the deviation amount of the cleaning device 2 from a predetermined reference position based on the detection results of the pipe detection unit 101, rather than the detection value of the proximity sensor 272. Figure 13 is a diagram showing an example of how to calculate the deviation amount based on the detection results of the pipe detection unit 101.

The configuration of the control device of this embodiment is the same as the configuration of control device 1 of embodiment 1 (see FIG. 1), with the only difference being the content of the processing performed by the deviation amount calculation unit 103. Therefore, the control device of this embodiment will also be referred to as control device 1. Furthermore, the cleaning device of this embodiment is the same as cleaning device 2 of embodiment 1, so the cleaning device of this embodiment will also be referred to as cleaning device 2. However, cleaning device 2 of this embodiment differs from cleaning device 2 of embodiment 1 in that it does not need to be equipped with a proximity sensor 272.

As described in embodiment 1, the pipe detection unit 101 detects pipes from images captured by the imaging device 271 attached to the cleaning device 2. In the example of Figure 13, four pipe contours are detected from image IMG6 captured by the imaging device 271.

More specifically, image IMG6 shows pipes PI11 to PI14, with pipes PI11 and PI14 located on the contact surface with the cleaning device 2 (top row), and pipes PI12 and PI13 located further back (second row). The right edge contour line of pipe PI11 and its upper end point p1, the right edge contour line of pipe PI12 and its upper end point p2, the left edge contour line of pipe PI13 and its upper end point p3, and the left edge contour line of pipe PI14 and its upper end point p4 have been detected.

When the image capture device 271 is positioned in the horizontal center of the cleaning device 2, the amount of deviation between the horizontal center of the area sandwiched between the uppermost pipes PI and PI14 (hereinafter referred to as the inter-pipe area) and the horizontal center of image IMG6 represents the amount of deviation of the cleaning device 2. Therefore, when calculating the amount of deviation, the deviation amount calculation unit 103 simply identifies the inter-pipe area from image IMG6.

To identify the inter-pipe region, the deviation amount calculation unit 103 uses the tilt angle identified by the angle identification unit 102 and the coordinate values of the upper end points p1 to p4 to identify the lower end points corresponding to the upper end points p1 to p4, and identifies a rectangular region defined by the upper and lower end points. The deviation amount calculation unit 103 then determines whether the identified rectangular region is an inter-pipe region.

For example, in IMG7 in Figure 13, the deviation amount calculation unit 103 identifies the bottom endpoint p6 corresponding to the top endpoint p4, and then identifies the bottom endpoint p5 corresponding to the top endpoint p3 adjacent to the top endpoint p4, thereby identifying a rectangular area defined by the four points p3, p4, p6, and p5.

Next, as shown in IMG8, the deviation amount calculation unit 103 masks the area outside the identified rectangular area with black and converts it to grayscale. The deviation amount calculation unit 103 then calculates a histogram of pixel values within the rectangular area, and determines that it is an inter-pipe area if the median is smaller than a threshold value (e.g., 100). Normally, the area containing the topmost pipes PI11 and PI14 has higher pixel values than the inter-pipe area, so the above determination makes it possible to distinguish between the area containing the topmost pipes PI11 and PI14 and the inter-pipe area. In this example, the rectangular area defined by the four points p3, p4, p6, and p5 is determined to be an inter-pipe area.

Similarly, in IMG9 in Figure 13, the deviation amount calculation unit 103 identifies the bottom endpoint p7 corresponding to the top endpoint p1. Here, after identifying the bottom endpoint p7, the deviation amount calculation unit 103 identifies the bottom endpoint p8 corresponding to the top endpoint p3 adjacent to the top endpoint p2 without identifying the bottom endpoint corresponding to the top endpoint p2, thereby identifying a rectangular area defined by the four points p1, p3, p8, and p7.

In this way, if the distance between a certain upper end point and its adjacent upper end point is equal to or less than a threshold, the deviation calculation unit 103 may identify the lower end point only for the certain upper end point, without identifying the lower end point for the adjacent upper end point. This is because it is unlikely that two adjacent end points are both the ends of a pipe.

The deviation amount calculation unit 103 identifies a rectangular area defined by four points p1, p3, p8, and p7, and then, as with IMG8, masks the area outside the identified rectangular area with black and converts it to grayscale (IMG10). The deviation amount calculation unit 103 then calculates a histogram of pixel values within the rectangular area, and determines that it is an inter-tube area if the median is smaller than a threshold. In this example, the rectangular area defined by four points p1, p3, p8, and p7 is determined to be an inter-tube area. This process is performed for all rectangular areas formed based on the detected upper endpoints.

Once the rectangular area based on the detected upper end point has been identified and whether or not that rectangular area is an inter-pipe area has been determined, the deviation amount calculation unit 103 integrates the identified inter-pipe areas. Specifically, the deviation amount calculation unit 103 identifies the coordinates of the vertices of the identified inter-pipe areas that are located closest to the upper left, the lower left, the lower right, and the upper right, and determines the area defined by these coordinates as the final inter-pipe area.

For example, in the example of Figure 13, the deviation calculation unit 103 identifies four vertices, p1, p7, p6, and p4, as shown in IMG11, from the vertices of the two inter-pipe regions identified as described above ("p3, p4, p6, p5" and "p1, p3, p8, p7"). As a result, the final inter-pipe region is identified as a rectangular region defined by the four points p1, p7, p6, and p4.

Finally, as shown in image IMG12 in Figure 13, the deviation amount calculation unit 103 determines the midpoint between points p1 and p4 and the midpoint between points p7 and p6, and calculates the distance Δy between line segment L11 connecting these midpoints and line segment L12 that bisects image IMG12 in the left-right direction as the deviation amount.

As described above, the control device 1 may be configured to include a deviation amount calculation unit 103 that calculates the amount of deviation of the cleaning device 2 from a reference position based on the detection results of the pipe detection unit 101, and the movement control unit 104 to control the movement of the cleaning device 2 based on the tilt angle identified by the angle identification unit 102 and the amount of deviation calculated by the deviation amount calculation unit 103.

The position of the pipe shown in the image captured by the camera 271 attached to the cleaning device 2 reflects the relative position of the cleaning device 2 and the pipe. Therefore, by detecting the pipe from this image, the amount of misalignment of the cleaning device 2 can be calculated. This amount of misalignment can then be used to control movement, allowing the cleaning device 2 to be aligned to a predetermined position relative to the pipe.

[Modification]
The execution entity of each process described in each of the above-described embodiments may be any entity and is not limited to the above-described examples. In other words, any device configuration may be used as long as the processes described in each of the above-described embodiments can be executed. For example, among the processes executed by the control device 1, the process of identifying the tilt angle, the process of calculating the amount of deviation, and the process of controlling the operation of the cleaning device 2 may be executed by different information processing devices. Furthermore, for example, a first control device 1 may perform the process of identifying the tilt angle and the process of controlling the operation of the cleaning device 2 based on the identified tilt angle, and a second control device 1 may perform the process of calculating the amount of deviation and the process of controlling the operation of the cleaning device 2 based on the calculated amount of deviation.

The control device 1 may also be mounted on the cleaning device 2. In other words, a cleaning device 2 equipped with the control device 1 and moving according to the control of the control device 1 is also included in the scope of the present invention. A cleaning device 2 equipped with the control device 1 can move independently to a predetermined position on the pipe.

[Software implementation example]
The functions of the control device 1 (hereinafter referred to as the "device") can be realized by a program for causing a computer to function as the device, and a program (control program) for causing a computer to function as each control block of the device (particularly each part included in the control unit 10).

In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The functions described in each of the above embodiments are realized by executing the program using this control device and storage device.

The above program may be stored non-transitory on one or more computer-readable storage media. These storage media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

Furthermore, some or all of the functions of each of the above control blocks can be realized by logic circuits. For example, integrated circuits incorporating logic circuits that function as each of the above control blocks are also included in the scope of the present invention. In addition, the functions of each of the above control blocks can also be realized by, for example, a quantum computer.

The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention.

(Additional Notes)
A control device according to one aspect of the present invention includes a pipe detection unit that detects the pipe from an image captured by an imaging device attached to a cleaning device that cleans the surface of the pipe, an angle identification unit that identifies the inclination angle of the pipe detected by the pipe detection unit, and a movement control unit that controls the movement of the cleaning device based on the inclination angle.

A cleaning device control method according to one aspect of the present invention is a cleaning device control method executed by a control device, and includes a pipe detection step of detecting the pipe from an image captured by an image capture device attached to the cleaning device that cleans the surface of the pipe, an angle identification step of identifying the inclination angle of the pipe detected in the pipe detection step, and a movement control step of controlling the movement of the cleaning device based on the inclination angle.

1 Control device 101 Pipe detection unit 102 Angle identification unit 103 Displacement amount calculation unit 104 Movement control unit 2 Cleaning device 271 Photography device 272 (272R, 272L) Proximity sensor

Claims (4)

a displacement amount calculation unit that calculates a displacement amount of the cleaning device from a predetermined reference position based on detection values of a pair of proximity sensors that detect the pipes and are attached to symmetrical positions on the left and right sides of the cleaning device to clean the pipes arranged in parallel;
a movement control unit that controls movement of the cleaning device based on the amount of deviation,
The pair of proximity sensors are arranged so that when one proximity sensor is located directly above the pipe, the other proximity sensor is located at a detection limit position of a pipe adjacent to the proximity sensor,
The deviation amount calculation unit calculates the deviation amount using an approximation formula that approximates the relationship between a difference between the detection values of the pair of proximity sensors and the deviation amount. A cleaning device equipped with the control device described in claim 1 and moving in accordance with the control of the control device. A cleaning device control method executed by a control device, comprising:
a displacement amount calculation step of calculating a displacement amount of the cleaning device from a predetermined reference position based on detection values of a pair of proximity sensors attached to symmetrical positions on the left and right sides of the cleaning device that cleans a plurality of pipes arranged in parallel;
a movement control step of controlling movement of the cleaning device based on the amount of deviation,
The pair of proximity sensors are arranged so that when one proximity sensor is located directly above the pipe, the other proximity sensor is located at a detection limit position of a pipe adjacent to the proximity sensor,
The method for controlling a cleaning device, wherein the deviation amount calculation step calculates the deviation amount using an approximation equation that approximates the relationship between a difference between the detection values of the pair of proximity sensors and the deviation amount. A control program for causing a computer to function as the control device described in claim 1, the control program causing the computer to function as the displacement amount calculation unit and the movement control unit.