JP6180496B2 - Terminal speed estimation method - Google Patents

Description

  The present invention relates to a terminal speed estimation method in a wireless communication system, and more particularly to a terminal speed estimation method using a Doppler spectrum.

  The wireless communication system assumes various uses of wireless communication by the user, from low data rate communication to high data rate communication, from still images to video streaming transmission, from stationary or walking state to high speed moving state, Alternatively, it provides wireless communication services in various forms such as indoor use and outdoor use.

  In addition to these types of usage, operators providing wireless communication systems have various conditions such as locations where services are provided, such as urban and rural areas, population distribution of users who use the services, and data volume (traffic density). We are building a system that takes into account. As part of the system construction, the operator should set the usage pattern and traffic density so that users can connect to the wireless communication system without delay at any location and time and receive satisfactory services. Consider the base station (or cell) layout design.

  However, with further diversification of usage forms and increase in traffic volume, it is difficult to provide sufficient services with a planar cell arrangement. For this reason, the construction of a hierarchical cell configuration that provides services that match various usage forms and traffic characteristics is being promoted by combining a plurality of types of cells. One example of such a cell configuration is a hierarchical cell configuration in which a minimal cell having a small cell size is superimposed (overlaid) in a macro cell having a large cell size. The micro cell is a micro cell, a pico cell, a femto cell, or the like. Such a micro cell is installed in a place where the traffic density in the macro cell is high, a place where high data rate communication is assumed, a place where radio wave reception from the macro cell base station is difficult, and the like.

  In the hierarchical configuration, in order to improve frequency utilization efficiency or suppress the frequency of handover, hierarchical selection control is required to appropriately select the macro cell and the minimum cell hierarchy according to the moving speed of the terminal.

  A terminal having a high moving speed is connected only to a macro cell and not connected to a very small cell. On the other hand, a terminal having a low moving speed is connected to a minimal cell when in a minimal cell, and is connected to a macro cell at a location where the minimal cell is not superimposed. By such hierarchical selection control, it is possible to prevent a terminal moving at high speed from frequently handing over to a very small cell, and to reduce communication interruption of user data accompanying an increase in handover control signaling. Further, by connecting a low-speed moving terminal to a very small cell, a data offload effect can be expected, and user throughput can be increased.

  In order to appropriately execute the hierarchy selection control, it is important to accurately measure the moving speed of the terminal and connect to the macro cell or the minimal cell based on the moving speed.

  One method for estimating the moving speed of a terminal is a method using a Doppler spectrum (Non-patent Document 1). Here, the Doppler spectrum is a power waveform on the frequency axis (Doppler frequency shift) obtained as a result of Fourier transform of the time variation of the amplitude value of the propagation path response. When the radio propagation environment between the terminal and the base station can be regarded as an all-around scattering model with a uniform arrival direction distribution of radio waves, the Doppler spectrum shows a peak at the maximum Doppler frequency. For this reason, the maximum Doppler frequency can be measured by detecting the peak of the spectrum, and the moving speed of the terminal can be calculated from the maximum Doppler frequency.

G-H. Park, D. Hong and C-E. Kang, `` A New Doppler Spread Estimation Using FFT '' IEICE Trans.Commun., Vol.E86-B, No.9, pp.2799-2803, Sept. 2003.

  However, since the amplitude value of the propagation path response includes noise and interference, the measured estimated value of the maximum Doppler frequency includes an error. Further, by performing Fourier transform using a limited number of amplitude values, distortion occurs in the Doppler spectrum and affects the estimated value.

  Furthermore, there may be a case where an accurate maximum Doppler frequency cannot be detected from the Doppler spectrum due to the radio propagation environment. For example, when the noise level instantaneously increases in a band outside the Doppler spectrum band, it is conceivable that the noise component is detected as an effective component of the Doppler spectrum. In this case, the Doppler spectrum width is widened, and the estimated speed value is calculated to be higher than the actual traveling speed.

  Further, in a wireless propagation environment in which there are few scattered wave components arriving from the traveling direction of the terminal and in the opposite direction, the Doppler spread becomes smaller than when there is a sufficient scattered wave component. As a result, the estimated speed value is calculated lower than the actual traveling speed.

  The present invention calculates a high-accuracy maximum Doppler frequency and estimates a moving speed with higher accuracy in consideration of the influence of errors and noise during Doppler spectrum measurement in estimating the moving speed of a terminal using the Doppler spectrum. An object is to provide a speed estimation method.

  The present invention relates to a method for estimating a moving speed of a terminal, the step of receiving a reference signal, the step of creating time variation data of a propagation path from the received reference signal, and the time variation data of the propagation path Multiplying FFT, calculating a Doppler spectrum from the output of the FFT, detecting a maximum Doppler frequency from the Doppler spectrum, determining an instantaneous moving speed of the terminal from the maximum Doppler frequency, and The step of determining the instantaneous movement speed of the terminal is executed a plurality of times to obtain a plurality of instantaneous movement speeds, and the weighted average is calculated by multiplying each of the obtained plurality of instantaneous movement speeds by a weighting coefficient. Thereby determining the moving speed of the terminal.

  In the moving speed estimation method according to the present invention, it is possible to estimate a more accurate moving speed in consideration of the influence of errors and noise when measuring the Doppler spectrum.

It is a figure which shows the example of a Doppler spectrum. It is a figure explaining the detection method of the maximum Doppler frequency. It is a figure which shows the flowchart of the estimation method of the maximum Doppler frequency. It is a figure explaining the influence of the noise in maximum Doppler frequency estimation. It is a figure explaining the Doppler spectrum of a condition with few scattered wave components. It is a figure which shows the Example of this invention. It is a figure which shows the flowchart of another estimation method of the maximum Doppler frequency. It is a figure explaining application of the threshold value to a Doppler spectrum. It is a figure which shows the flowchart of the other estimation method of the maximum Doppler frequency. It is a figure which shows the flowchart of another estimation method of the maximum Doppler frequency. It is a figure which shows the example of the apparatus for estimation of the maximum Doppler frequency.

  Hereinafter, a terminal speed estimation method using a Doppler spectrum will be described with reference to the drawings. However, it should be understood that the invention is not limited to the drawings or the embodiments described below.

  FIG. 1 shows an example of a Doppler spectrum. This example represents an ideal case where the radio propagation environment can be regarded as an all around scattering model. In the figure, the horizontal axis represents frequency, and the vertical axis represents relative power. However, since FIG. 1 shows the processing result after conversion to the baseband, it should be noted that the frequency 0 Hz on the horizontal axis actually corresponds to the carrier frequency. In the subsequent Doppler spectrum diagrams, the definition of the frequency on the horizontal axis is the same as in FIG. In this example, the maximum Doppler frequency is 100 Hz, and high power components occur at frequencies of 100 Hz and -100 Hz in the Doppler spectrum. Thus, the maximum Doppler frequency can be estimated by detecting the frequency of the component with high power from the Doppler spectrum calculated using the time fluctuation of the propagation path of the received signal.

FIG. 2 is a diagram simply illustrating a method for calculating the maximum Doppler frequency from the Doppler spectrum. On the negative side of the frequency, f (1) is the frequency at which the maximum power is obtained, and on the positive side, the frequency at which the maximum power is obtained is f (2). Under an ideal fading environment, a U-shaped spectrum is shown between f (1) and f (2), and f (1) and f (2) show the same absolute frequency value. The frequency absolute value is the maximum Doppler frequency f _D. That is, f (1) = − f _D and f (2) = f _D. The frequency between f (1) and f (2) is the Doppler spectrum width.

When the maximum Doppler frequency f _D is obtained, the moving speed ν of the terminal can be calculated by f _D × λ. Here, λ is the wavelength of the carrier frequency.

  However, in an actual wireless propagation environment, the arrival direction distribution around the terminal is not necessarily uniform, and the received power is not uniform in all directions. In addition, due to the influence of noise, the Doppler spectrum is not a smooth U-shape.

  The Doppler spectrum is calculated by performing FFT on the received signal power data. However, since FFT performs Fourier transformation with a finite number of points, both ends of the original time variation data become discontinuous. The resulting spectrum is distorted. The occurrence of this distortion means that a component that does not originally exist appears in the Doppler spectrum obtained by the FFT, which causes an error in determining the maximum Doppler frequency.

  First, a method for calculating the maximum Doppler frequency with higher accuracy in consideration of these effects will be described.

  FIG. 3 is a diagram illustrating a flowchart of a method for estimating the maximum Doppler frequency using the Doppler spectrum. In S10, a reference signal is received from a wirelessly connected base station or terminal. Although the reference signal is used here, it is also known as a pilot signal or a beacon depending on the system, and is a known signal used for propagation path estimation. In S11, the received reference signal is converted into a baseband signal, demodulated, and then received power is measured to create time variation data of the propagation path of the received signal. The time variation data of the propagation path is stored in the buffer for every fixed section.

  Since the received signal varies in proportion to the moving speed of the terminal, a Doppler spectrum can be obtained by performing Fourier transform on the time variation of the propagation path of the received signal. In S12, fast time Fourier transform (FFT) is applied to the time variation data of the propagation path stored for each fixed section. Here, the fixed interval corresponds to the number of points in the FFT. In S13, a Doppler spectrum is calculated from the output of the FFT.

In S14 of FIG. 3, the component having the maximum power in the region where the frequency of the Doppler spectrum is positive is found, and the frequency f (P _{max +} ) is detected. As a search method for the maximum power component, for example, the power P (f = 0) at the frequency 0 Hz is compared with the power P (f = Δf) at the next frequency Δf Hz, and the high power component is selected. Further, the power P (f = 2 × Δf) at the next frequency 2 × Δf Hz is compared, and a higher power component is selected. In this way, the power of each component in the positive frequency region is compared in order, the component with the highest power is finally selected, and the frequency is defined as f (P _{max +} ).

In S15, the frequency f (P _max− ) at which the maximum power is obtained in the region where the frequency of the Doppler spectrum is negative is detected by the same method as in S14.

In S16, the maximum Doppler frequency is calculated using the two detected frequencies. Here, of the frequency f (P _{max +} ) detected on the positive side of the frequency and the frequency f (P _max− ) detected on the negative side, the larger absolute value is set as the maximum Doppler frequency f _D. That is, f _D = max (| f (P _{max +} ) |, | f (P _max− ) |).

  When a radio wave having strong received power comes from a specific direction at the time of signal reception, a strong power component may be generated between the positive and negative maximum Doppler frequencies, that is, both peaks of the Doppler spectrum. For this reason, the case where the frequency which shows maximum electric power does not correspond with the maximum Doppler frequency can be considered. By selecting the larger absolute value of the positive and negative maximum power frequencies detected in S14 and S15, it is possible to determine a more accurate maximum Doppler frequency.

  Hereinafter, a method of calculating the moving speed of the terminal from the calculated maximum Doppler frequency and then calculating the moving speed with higher accuracy will be described. The estimated value of the moving speed obtained from the measurement of one maximum Doppler frequency is defined as the instantaneous moving speed.

First, a state where an error occurs due to the signal propagation environment in the calculation of the maximum Doppler frequency will be described. FIG. 4 is a diagram simply illustrating an example in which a noise component is detected as an effective component. In a frequency range higher than the positive maximum Doppler frequency (f _D ) and / or a frequency range lower than the negative maximum Doppler frequency (−f _D ), the noise level may increase instantaneously. By detecting this noise component as an effective component of the Doppler spectrum, the spectrum width becomes wider than the actual spectrum width. As a result, the calculated speed estimated value becomes higher than the actual terminal moving speed. Since noise outside the band of the Doppler spectrum is regarded as an effective component, the error of the speed estimation value can be very large.

  FIG. 4 shows an example in which noise components are detected as effective components on both sides of a frequency region higher than the positive maximum Doppler frequency and a frequency region lower than the negative maximum Doppler frequency, but only one side is detected. A situation also arises.

FIG. 5 shows a Doppler spectrum in the case where there are few scattered wave components coming from the traveling direction and the direction opposite to the traveling direction when the terminal moves. If less scattered wave component, the spectral width is smaller than the actual spectral width (-f _D ~f _D). As a result, the estimated speed value is calculated low.

  Even if the signal propagation environment is ideal, that is, a case where it can be regarded as an all-around scattering model, the estimated speed value includes variations due to the influence of a finite measurement time, noise, and the like.

  A method for estimating the moving speed more accurately in consideration of the influence of the propagation environment and the measurement error will be described below.

The measurement of the instantaneous moving speed using the Doppler spectrum is performed a plurality of times, and the influence of the error is suppressed by multiplying each instantaneous moving speed by a weighting coefficient and taking a weighted average. Speed measurement is performed N times, and each measured value is set to v ₁ , v ₂ ,... V _N km / h. The calculation of the instantaneous movement speed is performed by the method according to the flowchart of FIG. 3, but is not limited to this method. A simple example of the weighted average is shown in FIG.

ここで、w_i (i=1〜N)は、重み付け係数である。 The weighted average value is calculated as follows.

Here, w _i (i = 1 to N) is a weighting coefficient.

In the above equation (1), the simplest weighting method is to set all the weighting coefficients w _i (i = 1 to N) to 1. That is, averaging is performed. When the instantaneous movement speed value includes a variation due to a measurement error, the influence of the variation can be suppressed by averaging. Here, the number of trials N is assumed to be several times. In order to calculate one Doppler spectrum, the time for measuring the time variation data of the propagation path is assumed to be 256 msec. For this reason, when N is set to several times, the moving speed of the terminal itself changes greatly while calculating the weighted average, and the estimated speed value is hardly affected. However, these conditions are not limited to the values shown here because they are related to the processing capability of the apparatus, the amount of memory, and the like.

Next, a case where a large noise component is instantaneously generated as shown in FIG. 4 and a large error is included in the measured value of the instantaneous moving speed is considered. By measuring such a measured value including a large error as an abnormal value and excluding this abnormal value, it is possible to estimate the speed more accurately. If an abnormal value is detected, it can be excluded by setting the weighting coefficient w _i of the measured value to 0 in equation (1).

An example of a method for excluding abnormal values is shown below. The moving speed is measured N times, and each measured value is set to v ₁ , v ₂ ,... V _N km / h. The m-th measurement value v _m is compared with other measurement values v _n (n ≠ m), and it is determined whether or not the measurement value v _m has a value significantly different from the other measurement values.

速度差のしきい値を超える割合P_mが所定しきい値Pより大きい場合（P_m＞P）、ｍ番目の測定値は異常値と判断する。 Calculate the absolute difference | v _m −v _n | between the measured value v _m and the other measured values v _n , as shown in the following equation, and the absolute difference represents the speed difference threshold (v _thresh ). Calculate the percentage exceeding.

When the ratio P _m exceeding the speed difference threshold is larger than the predetermined threshold P (P _m > P), the m-th measured value is determined to be an abnormal value.

Specific examples are shown below. Assume that three speed measurements were made (N = 3). The measured values are v ₁ = 38 km / h, v ₂ = 40 km / h, and v ₃ = 60 km / h. The following table calculates the absolute difference between each measured value.

When the speed difference threshold (v _thresh ) is 5 km / h, P ₁ = 0.5, P ₂ = 0.5, and P ₃ = 1.0. If the predetermined threshold value P is 0.5, only P ₃ exceeds the predetermined threshold value, so the measured value of m = 3 is determined to be abnormal, and the weighting coefficient w ₃ = 0 in the weighting calculation is set. Excluded from the weighting operation.

  In the above embodiment, it is assumed that the number of instantaneous movement speed samples is as small as several. This is because, for example, when the terminal is on the vehicle, in the speed estimation, it responds to a rapid speed change due to the stop or start of the car. However, if the car is traveling on a highway and a stable speed continues for a long section, or if the terminal is in the hand of a walking person and the walking speed is constant, increasing the number of samples of the instantaneous movement speed will make it more accurate. Speed estimation is possible.

  Hereinafter, a method for excluding abnormal values when the number of samples of the instantaneous movement speed is increased (N is several tens or more) will be described. Here, the cumulative distribution of the instantaneous moving speed is calculated, and the instantaneous moving speed at which the cumulative distribution probability becomes a certain value or more is excluded as an abnormal value. When the cumulative distribution of the instantaneous moving speed based on the Doppler frequency is calculated, the probability that the instantaneous moving speed is measured lower than the actual value is relatively small. This is because the probability of occurrence of a radio propagation situation as shown in FIG. 5, that is, a situation in which there are few scattered wave components from the traveling direction of the terminal and the opposite direction is low. In addition, the error amount of the measured value at this time is small. On the other hand, the situation where the noise component outside the Doppler spectrum increases instantaneously and the instantaneous moving speed is measured higher than the actual value occurs at a relatively high frequency. And the error amount of the measured value in this case may become large. For this reason, here, only measured values having a cumulative distribution probability of a certain value or more are excluded.

  An example of a specific processing method will be described. Let N be the number of samples of the measured instantaneous movement speed. Here, N is several tens or more, and is a sufficiently large number to obtain a cumulative distribution. In the cumulative distribution probability, a measured value having a probability of Y% or more is determined as an abnormal value. For example, when Y = 95 (%), (0.05 × N) samples are determined to be abnormal values.

The measured instantaneous moving speeds v ₁ , v ₂ ,... V _N km / h are sorted and arranged in order of increasing speed. A sample (0.05 × N) with a high speed is determined as an abnormal value, and its weight w _{i is set} to zero. However, (0.05 × N) is an integer.

  As an example of specific processing, u (0.05 × N, 2) buffers are provided in the memory of the apparatus. u (0, 0) is the velocity value of the sample with the highest instantaneous movement speed, u (0, 1) is the sample number (1 to N) of that sample, and u (1, 0) is the second The velocity value of the sample with a high velocity, u (1, 1) is the sample number of the sample, ... u (0.05 × N-1, 0) is the sample with the (0.05 × N) -th highest velocity The sample number of the sample is recorded in the velocity value u (0.05 × N−1, 1). When a new sample is measured, the velocity value in the buffer is compared with the velocity value in the buffer (u (0, 0) to u (0.05 × N-1, 0)) and is higher than the velocity value in the buffer The velocity value and sample number of the sample are newly recorded. When the measurement is completed, the weighting coefficient is set to 0 for the (0.05 × N) sample number samples recorded in the buffer. An averaging process is performed after removing (0.05 × N) samples.

  By excluding the instantaneous moving speed of 95% or more of the cumulative distribution as an abnormal value, it is possible to estimate the speed with higher accuracy. Here, it is 95% or more, but is not limited to this value. Also, cumulative distributions below a certain value (for example, 5% or less) can be excluded. Further, it can be set according to the signal propagation environment and the speed of the terminal.

  If the estimated value of the moving speed is calculated to be higher than the actual value, it is determined that the moving speed is moving faster, and the possibility of handover from the minimal cell to the macro cell increases. Handover to the macro cell consumes valuable radio resources in the base station of the macro cell. In addition, the data offload effect to the tiny cell cannot be expected. . For this reason, it is advantageous to improve the frequency utilization efficiency to make the abnormal value more than a certain probability of the cumulative distribution.

  The example of the movement speed estimation when the number of samples of the instantaneous movement speed is small and large is described. It is also possible to combine these two. For example, speed estimation with a small number of samples is repeatedly performed, and the moving speed measurement value of the sample used for estimation is stored. If multiple speed estimations show similar speeds, determine that the terminal is moving at a constant speed, determine the cumulative distribution from the stored speed measurements, and determine a more accurate average value Can do.

  As a further embodiment, a method using a forgetting factor will be described. The weighting is increased for the recent instantaneous moving speed measurement result, and the weighting is decreased for the old result. In an environment where the speed of the terminal changes, the weight of the latest measurement result becomes large, so that more accurate speed estimation can be performed.

  When instantaneous movement speed is measured at each time interval T, the weighting coefficient for each sample is, for example, w (t = 0) = 1, w (t = -T) = 1-α, w (t =- 2T) = 1-2α... Here, 0 <α <1 and 0 ≦ w ≦ 1 are set. However, α need not be a constant value. For example, the more recent velocity measurement result can be strongly reflected by setting the α value for an older sample larger.

  It is also conceivable to use a forgetting factor after excluding the abnormal values described above. In this case, the weighting coefficient of the sample determined to be an abnormal value is set to 0, and the forgetting coefficient is reflected in the weighting coefficients of the remaining samples.

  In the following, another method for detecting the maximum Doppler frequency will be described. The above-described weighted average method can also be applied to the instantaneous moving speed calculated using the maximum Doppler frequency detected in these methods.

FIG. 7 is a diagram illustrating a flowchart of another method for calculating the maximum Doppler frequency. In FIG. 7, the processing (S10 to S13) until the Doppler spectrum is calculated is the same as the method shown in FIG. In S24, the frequency f (P _max1 ) of the component having the first highest power is detected. Subsequently, in S25, the frequency f (P _max2 ) of the component having the second highest power is detected. In S26, the maximum Doppler frequency is calculated. Here, the spectrum width is calculated from f (P _max1 ) and f (P _max2 ), and half of the spectrum width is set as the maximum Doppler frequency. That is, let | f (P _max1 ) −f (P _max2 ) | / 2 be the maximum Doppler frequency.

  The instantaneous moving speed of the terminal is calculated from the maximum Doppler frequency detected by the method shown in FIG. The above weighted average method can be applied to this instantaneous moving speed.

  In another method, a method will be described in which a threshold is set for the Doppler spectrum and the maximum Doppler frequency is calculated from the maximum frequency component and the minimum frequency component of the active components that are equal to or greater than the threshold.

FIG. 8 is a diagram showing an example of threshold setting in the Doppler spectrum. Here, the noise power N ₀ [dBm] is calculated, and a value obtained by adding ΔN [dB] calculated from the noise variance is used as a threshold value. As shown in FIG. 8, by using N ₀ + ΔN [dBm] as a threshold value, noise components outside the Doppler spectrum can be excluded. And the effective component which becomes more than a threshold value becomes a component with some high electric power containing a direct wave component. Among these active components, the component having the maximum frequency (ie, the 100 Hz component indicated by “A” in FIG. 8) and the component having the minimum frequency (ie, indicated by “B” in FIG. 8) By selecting the 100 Hz component), the maximum Doppler frequency can be determined.

FIG. 9 is a flowchart showing still another method for calculating the maximum Doppler frequency. The processing (S10 to S13) until the Doppler spectrum is calculated is the same as the method shown in FIG. In S34, a threshold value of the spectrum is set, and an active component that is equal to or higher than the threshold value is detected. In S35, the effective component E _{f_max} having the maximum frequency is selected from the effective components, and the frequency f (E _{f_max} ) is detected. In S36, the effective component E _{f_min} having the minimum frequency is selected from the effective components, and the frequency f (E _{f_min} ) is detected.

In S37, the maximum Doppler frequency is calculated using the two detected frequencies. Here, of the frequency f (E _{f_max} ) and the frequency f (E _{f_min} ), the larger absolute value is set as the maximum Doppler frequency f _D. That is, f _D = max (| f (E _{f_max} ) |, | f (E _{f_min} ) |).

  In this way, by setting a threshold value and extracting an effective component that is equal to or higher than the threshold value, it is possible to exclude only the spectrum due to the noise component and select only the spectrum due to the received signal. Furthermore, by providing an appropriate threshold value, it is possible to reduce false detection due to noise components and estimate a more accurate maximum Doppler frequency.

In the above example, the threshold value is set to a value obtained by adding ΔN to the noise power N ₀ , but is not limited to this. For example, the threshold value can be made variable according to the radio propagation status such as the measured SNR.

  The instantaneous moving speed of the terminal is calculated from the maximum Doppler frequency detected by the method shown in FIG. The above weighted average method can be applied to this instantaneous moving speed.

FIG. 10 is a flowchart showing another method for calculating the maximum Doppler frequency using a threshold value. The processing (S10 to S13) until the Doppler spectrum is calculated is the same as the method shown in FIG. In S44, a threshold value of the spectrum is set, and an active component exceeding the threshold value is detected in the spectrum. In S45, the effective component E _{f_max} having the maximum frequency is selected from the effective components, and the frequency f (E _{f_max} ) is detected. In S46, the effective component E _{f_min} having the minimum frequency is selected from the effective components, and the frequency f (E _{f_min} ) is detected.

In S47, the maximum Doppler frequency is calculated using the two detected frequencies. Here, the spectrum width is calculated from f (E _{f_max} ) and f (E _{f_min} ), and half of the spectrum width is set as the maximum Doppler frequency. That is, let | f (E _{f_max} ) −f (E _{f_min} ) | / 2 be the maximum Doppler frequency.

  The instantaneous moving speed of the terminal is calculated from the maximum Doppler frequency detected by the method shown in FIG. The above weighted average method can be applied to this instantaneous moving speed.

  FIG. 11 is a block diagram for briefly explaining the apparatus configuration of the present invention. Such a configuration can be installed on either the base station side or the terminal side, and enables measurement of the maximum Doppler frequency in both stations.

  In FIG. 11, 50 is a signal receiving unit, 51 is a received signal power measuring unit, 52 is an FFT, 53 is a signal selecting unit, 54 is a maximum Doppler frequency calculating unit, and 55 is a terminal speed estimating unit. The signal receiving unit 50 receives a signal from a transmitting station that performs wireless communication, and extracts a reference signal from the received signal. The method of inserting the reference signal is defined by the system. The received signal power measurement unit 51 converts the received reference signal into a baseband signal, demodulates it, measures the received power, and generates time variation data of the propagation path of the received signal. The generated time variation data of the propagation path is stored in the buffer. The received signal power measurement unit 51 can also be implemented as part of propagation path estimation. The FFT 52 multiplies the time variation data of the propagation path by the FFT to calculate a Doppler spectrum. The signal selection unit 53 selects a component to be used for calculating the maximum Doppler frequency from the Doppler spectrum. In the example of FIG. 7, the signal selection unit 53 selects the first highest power component and the second highest power component. In the example of FIGS. 9 and 10, the signal selection unit 53 sets a threshold value, and selects the maximum frequency component and the minimum frequency component in the effective components. The maximum Doppler frequency calculation unit 54 calculates the maximum Doppler frequency from the frequency of the component selected by the signal selection unit 53. 7 and 10, the maximum Doppler frequency calculation unit 54 calculates the width of the spectrum and sets 1/2 of the spectrum as the maximum Doppler frequency. In the example of FIG. 9, the larger absolute value is the maximum Doppler frequency. The terminal speed estimation unit 55 calculates the moving speed of the terminal from the detected maximum Doppler frequency.

50 Signal receiver
51 Received signal power measurement section
52 FFT
53 Signal selector
54 Maximum Doppler frequency calculator
55 Terminal speed estimator

Claims (2)

A method for estimating the moving speed of a terminal,
Receiving a reference signal; and
Creating propagation time variation data from the received reference signal;
Multiplying the time variation data of the propagation path by FFT;
Calculating a Doppler spectrum from the output of the FFT;
Detecting a maximum Doppler frequency from the Doppler spectrum;
Determining an instantaneous moving speed of the terminal from the maximum Doppler frequency;
Obtaining a plurality of instantaneous movement speeds by executing the step of determining the instantaneous movement speed of the terminal a plurality of times;
A step of determining whether each of the plurality of instantaneous moving speeds is a normal value or an abnormal value, wherein a weighting coefficient for the instantaneous moving speed determined to be an abnormal value is set to 0, and an instantaneous value determined to be a normal value A step of setting the weighting coefficient for the moving speed to 1, and
Determining a moving speed of the terminal by multiplying each of the obtained plurality of instantaneous moving speeds by a weighting coefficient and calculating a weighted average;
I have a,
The step of determining whether the normal value or the abnormal value is:
Calculating an absolute value of a difference between one of the plurality of instantaneous moving speeds and another instantaneous moving speed;
Determining whether the calculated absolute value of each difference exceeds a first threshold;
Calculating a probability that an absolute value of a difference with respect to the one instantaneous moving speed exceeds the first threshold value from the result of the determination;
When the probability of exceeding the first threshold is less than or equal to a second threshold, the one instantaneous moving speed is regarded as a normal value;
When the probability of exceeding the first threshold exceeds the second threshold, the one instantaneous moving speed is regarded as an abnormal value;
A method for estimating a moving speed. The method according to claim 1 , wherein a forgetting factor is applied to the weighting factor.

Images