Multiple propagation wave parameter measuring method and apparatus and machine-readable recording medium recording multiple propagation wave parameter measuring program

[0002] The present invention relates to a multiple propagation wave parameter measuring method and apparatus for measuring various parameters of a multiple propagation wave, and a machine-readable recording medium recording a multiple propagation wave parameter measuring program and, more particularly, to a multiple propagation wave parameter measuring method and apparatus for measuring as parameters an arrival direction of a transmission wave radiated from a transmitting means and a change in delay time of the transmission wave from the transmitting means to a receiving means, and a machine-readable recording medium recording a multiple propagation wave parameter measuring program.

[0003] FIG. 10 shows the arrangement of a conventional multiple propagation wave parameter measuring apparatus. The multiple propagation wave parameter measuring apparatus shown in FIG. 10 comprises a transmitting device 101 for radiating transmission waves 104 in air and a receiving device 102 for receiving the transmission waves 104 and measuring various parameters.

[0004] When the transmission waves 104 are radiated from the transmitting device 101, the receiving device 102 directly receives some of the transmission waves 104. Reflecting members 103 (e.g., buildings) having the function of reflecting an electric wave reflect some of the transmission waves 104, and the receiving device 102 receives the reflected transmission waves. As described above, the transmission waves 104 arrive at the receiving device 102 via various routes. Therefore, the receiving device 102 receives them as a multiple propagation wave.

[0005] The receiving device 102 measures the arrival directions of the transmission waves 104 and their delay times from the transmitting device 101 to the receiving device 102 as follows.

[0006] An evaluation function is generated to estimate the arrival directions and delay times. This evaluation function is a function of an angle (arrival direction) and time (delay time).

[0007] The angular range of 0 to Θ and the time range of 0 to t in the evaluation function are calculated at intervals of ΔΘ and Δt, respectively. The peaks of the calculation values are found to estimate the arrival directions and delay times.

[0008] The delay times of two signals received at different timings are estimated by the above method, and the difference between the delay times is calculated to obtain a change in delay time.

[0009] To measure the change in delay time with high precision, a sync signal free from an influence of noise or the like must be used. For this purpose, in the multiple propagation wave parameter measuring apparatus shown in FIG. 10, to supply a sync signal from the transmitting device 101 to the receiving device 102 via a cable 105, the cable 105 connects the transmitting and receiving devices 101 and 102.

[0010] In the conventional multiple propagation wave parameter measuring apparatus shown in FIG. 10, a two-dimensional search is performed to estimate the two parameters, i.e., “arrival direction” and “change in delay time”. This method, however, is time-consuming in signal processing, and a measurement result cannot be immediately obtained.

[0011] In the conventional multiple propagation wave parameter measuring apparatus, the cable 105 connects the transmitting and receiving devices 101 and 102. For this reason, the conventional multiple propagation wave parameter measuring apparatus is not suitable for measuring a change in delay time by moving the transmitting device 101 outdoors.

[0012] The present invention has been made in consideration of the conventional problems described above, and has as its object to shorten the measurement time for two parameters, i.e., “arrival direction” and “change in delay time”.

[0013] It is another object of the present invention to allow measuring a change in delay time by moving at least one of transmitting and receiving sides.

[0014] In order to achieve the above objects of the present invention, there is provided a multiple propagation wave parameter measuring method comprising the steps of radiating transmission waves from transmitting means into an outer space, receiving the transmission waves by receiving means as a multiple propagation wave, measuring arrival directions of the transmission waves on the basis of reception signals output from the receiving means, defining the transmission wave arriving from one direction of the measured arrival directions as a desired wave and the transmission waves arriving from remaining directions as unnecessary waves, calculating a weight with which a reception power ratio of the desired wave to the unnecessary waves becomes maximum, and multiplying the reception signals with the weight to extract the reception signal in which the unnecessary waves are suppressed, and measuring a change in delay time of the desired wave from the transmitting means to the receiving means on the basis of the reception signal in which the unnecessary waves are suppressed.

[0015] FIG. 1 is a view showing the arrangement of a multiple propagation wave parameter measuring apparatus according to the first embodiment of the present invention;

[0016] FIG. 2 is a block diagram showing the arrangement of a receiving device shown in FIG. 1;

[0017] FIG. 3 is a view for explaining moving space averaging;

[0018] FIGS. 4A and 4B are views for explaining reception signals sampled by A/D converters;

[0019] FIG. 5 is a flow chart showing the operation flow of a signal processor shown in FIG. 2;

[0020] FIG. 6 is a flow chart showing the operation flow of an arrival direction measuring unit shown in FIG. 2;

[0021] FIG. 7 is a flow chart showing the operation flow of a desired wave extracting unit shown in FIG. 2;

[0022] FIG. 8 is a flow chart showing the operation flow of a delay time change measuring unit shown in FIG. 2;

[0023] FIG. 9 is a block diagram showing the arrangement of a signal processor in a multiple propagation wave parameter measuring apparatus according to the second embodiment of the present invention;

[0024] FIG. 10 is a view showing a conventional multiple propagation wave parameter measuring apparatus.

[0025] The preferred embodiments of the present invention will be described with reference to the accompanying drawings.

[0026] FIG. 1 shows a multiple propagation wave parameter measuring apparatus according to the first embodiment of the present invention. The multiple propagation wave parameter measuring apparatus shown in FIG. 1 comprises a transmitting device 1 serving as a transmitting means and a receiving device 2.

[0027] The transmitting device 1 radiates transmission waves 4 in air (outer space). The transmission waves 4 radiated from the transmitting device 4 form a periodic spread spectrum signal.

[0028] The receiving device 2 receives the transmission waves 4 as a multiple propagation wave and measures the arrival directions of the transmission waves 4 and their changes in delay times from the transmitting device 1 to the receiving device 2.

[0029] The receiving device 2 is fixed at a stationary position, and the transmitting device 1 is mounted on a movable carrier (not shown) and movable.

[0030] FIG. 2 shows the arrangement of the receiving device 2 shown in FIG. 1.

[0031] The receiving device 2 shown in FIG. 2 comprises M antenna elements 11 to 1M (M is an integer of two or more), M receivers 21 to 2M respectively connected to the outputs of the antenna elements 11 to 1M, M A/D (analog/digital) converters 31 to 3M respectively connected to the outputs of the receivers 21 to 2M, and a signal processor 40 connected to the outputs of the A/D converters 31 to 3M.

[0032] The signal processor 40 comprises an FFT (Fast Fourier Transform) processor 41 connected to the outputs of the A/D converters 31 to 3M, an arrival direction measuring unit 42 connected to the output of the FFT processor 41, a desired wave extracting unit 43 connected to the outputs of the FFT processor 41 and arrival direction measuring unit 42, a storage unit 44, and a delay time change measuring unit 45 connected to the outputs of the desired wave extracting unit 43 and storage unit 44.

[0033] The antenna elements 11 to 1M are aligned in one direction to constitute a linear antenna array 10 serving as a receiving means. Antenna elements 11 to 1M receive the transmission waves 4 and supply the. reception signals to the receivers 21 to 2M.

[0034] The receivers 21 to 2M have the function of converting the reception signals output from the antenna elements 11 to 1M into band-limited IFs (intermediate frequencies) which can be digitally processed.

[0035] The A/D converters 31 to 3M sample the reception signals output from the receivers 21 to 2M and convert the sampled reception signals into digital signals. The A/D converters 31 to 3M are set to sample the input signals for the same period of time at intervals of integer multiples of the periods of the transmission waves 4.

[0036] The FFT processor 41 in the signal processor 40 has the function of performing the FFT process of the reception signals output from the A/D converters 31 to 3M and analyze the frequencies of the FFT signals.

[0037] The arrival direction measuring unit 42 has the function of processing the reception signals output from the FFT processor 41 in accordance with moving space averaging and a MUSIC (Multiple Signal Classification) method and measuring arrival directions θ of the transmission waves 4.

[0038] The desired wave extracting unit 43 has the function of extracting only a reception signal having a desired wave from the reception signals output from the FFT processor 41. The desired wave is the transmission wave 4 arriving from one of the arrival directions measured by the arrival direction measuring unit. The transmission wave 4 arriving from another direction is called an unnecessary wave.

[0039] For example, if the transmission wave 4 arriving from the direction θ₁is defined as a desired wave, the transmission wave 4 arriving from a direction θ₂belongs to unnecessary waves. If the transmission wave 4 arriving from the direction θ₂is defined as a desired wave, the transmission wave 4 arriving from the direction θ₁belongs to unnecessary waves.

[0040] The storage unit 44 serves as a storage means pre-storing frequency spectrum data of the transmission waves 4 at a predetermined time.

[0041] The delay time change measuring unit 45 has the function of measuring a change Δτ in delay time of the desired wave by the moving space averaging and MUSIC method using the reception signal output from the desired wave extracting unit 43.

[0042] An arrival direction measuring algorithm for allowing the arrival direction measuring unit 42 to measure the arrival direction of the transmission wave 4 will be described below.

[0043] Reception signals x₁, x₂, . . . , X_M(complex numbers) received by the linear antenna array 10 made of the M antenna elements 11 to 1M are linear coupling results of D arrival signals F₁, F₂, . . . , F_D(complex numbers) and internal noise components (to be referred to as noise components hereinafter) n₁, n₂, . . . , n_M(complex numbers) of the M receivers 21 to 2M. The reception signals can be expressed by a matrix as follows:

[0044] X=AF+N  (1)

[0045] for

[0046] X=[x₁, x₂, . . . , x_M]^T  (2)

[0047] A=[a(θ₁), a(θ₂), . . . , a(θ_D)]  (3)

[0048] F=[F₁, F₂, . . . , F_D]^T  (4)

[0049] N=[n₁, n₂, . . . , n_M]^T  (5)

[0050] A matrix A represented by equation (3) is called an array manifold, and its column vector a (θ) is a response of the linear antenna array 10 to the direction θ and called a steering vector. This steering vector a (θ) is expressed as follows using responses a₁(θ), a₂(θ), . . . , a_M(θ) (complex numbers) to the directions θ of the M antenna elements 11 to 11M:

[0051] a(θ)=[a₁(θ) a₂(θ), . . . , a_M(θ)]^T  (6)

[0052] where T is the transposed matrix.

[0053] If the covariance matrix of equation (1) is defined as S, the following equation is obtained:

[0054] S_MM={overscore (XX  (7)

[0055] where H is the complex conjugate transposed matrix, and an overline indicates the average.

[0056] Assume that the arrival signals F₁to F_Dare not correlated with the noise components n₁to n_M, that the noise components n₁to n_Mare not correlated with each other, and that the M noise powers are equal to each other, i.e., σ²for descriptive convenience. The following equation is defined:

[0057] S_MM=A{overscore (FFA^H+σ²I=APA^H+σ²I  (8)

[0058] The following matrix is called a signal correlation matrix:

[0059] P={overscore (FF  (9)

[0060] where I is the unit matrix.

[0061] Since the D arrival signals F₁to F_Dare correlated with each other, the moving space averaging is performed for the covariance matrix S_MM.

[0062] FIG. 3 explains moving space averaging. According to the moving space averaging, the covariance matrix S_MMobtained from the linear antenna array 10 made of M antenna elements is divided into sub-array matrices S_mmin units of m elements, as shown in FIG. 3, and the sub-array matrices S_mmare added to obtain a sum. The moving space averaging is performed for the covariance matrix S_MMto suppress correlation among the arrival signals F₁to F_D.

[0063] The covariance matrix S_MMundergone moving space averaging is defined as S to obtain the following equation:

[0064] S=APA^H+σ²I  (10)

[0065] At this time, the steering vector a(θ) is given as follows:

[0066] a(θ)=[a₁(θ), a₂(θ), . . . a_m(θ)]^T  (11)

[0067] for m>D

[0068] The direction of an arrival signal will be estimated using the MUSIC method.

[0069] The covariance matrix S is decomposed into m eigenvalues λ₁, λ₂, . . . , λ_D, λ_D+1, . . . , λ_m. In this case, the following condition holds:

[0070] λ₁≧λ₂≧ . . . ≧λ_D  (12)

[0071] λ₁to λ_Dare eigenvalues of the arrival signals, and λ_D+1to λ_m(=σ²) are the eigenvalues of the noise components.

[0072] When the eigenvector for the eigenvalue of the noise component of the covariance matrix S is given as E_N, the following equation holds:

[0073] SE_N=σ²E_N  (13)

[0074] A substitution of equation (10) into equation (13) yields:

[0075] (APA^H+σ²I)E_N  (14)

[0076] When the arrival signals F₁to F_Dare not perfectly correlated with each other, and all the arrival directions are different, the following equation holds:

[0077] A^HE_N=0  (15)

[0078] The peak is searched using the following equation as an evaluation function to estimate the direction of the arrival signal:

[0079] P_MU(θ)=1/(a^H(θ)E_NE_N^Ha(θ))  (16)

[0080] A desired wave extraction algorithm for causing the desired wave extracting unit 43 to extract only a reception signal having a desired wave from the reception signals will be described below.

[0081] Assume that the transmission waves 4 arrive from the directions θ₁, θ₂, . . . , θ_D, and that only the transmission signal 4 arriving from the direction θ₁is extracted. In this case, the transmission wave 4 arriving from the direction θ₁is the desired wave, while the transmission waves 4 arriving from the remaining directions θ₂to θ_D, are unnecessary waves.

[0082] To extract only the reception signal having the desired wave from all the reception signals, a weight with which the reception power ratio of the desired wave to the unnecessary waves becomes maximum is obtained. The weight is multiplied with the reception signal to extract a reception signal in which the unnecessary waves are suppressed. That is, the weight is so obtained as to form null components in the directions θ₂to θ_Dto eliminate the unnecessary waves and pass only the desired wave arriving from the direction θ₁.

[0083] This will be described in more detail.

[0084] A reception signal X(t) contains a desired wave component X_d(t), an unnecessary component X_i(t), and a noise component X_n(t) of the receivers 21 to 2M. That is, X(t) is expressed as follows:

[0085] X(t)=X_d(t)+X_i(t)+X_n(t)  (17)

[0086] Assuming the covariance matrix of X_i(t)+X_n(t) as R, the following equation holds:

[0087] R=E{[X_i(t)+X_n(t)]*(X_i(t)+X_n(t)]^T}  (18)

[0088] where E{*} is the ensemble average, and * is the complex conjugate.

[0089] If the steering vector of the unnecessary wave is defined as A_i, equation (18) is equivalent to the following equation:

[0090] R={overscore (A²I  (19)

[0091] In addition, if the steering vector of the desired wave is defined as A_d, the weight vector W is defined as follows using the inverse matrix of the matrix R and the vector A_d:

[0092] W=αR⁻¹A_d*  (20)

[0093] where α is an arbitrary constant excluding zero.

[0094] When the reception signal X(t) containing all the components shown in equation (17) is multiplied with this weight vector W to obtain W*X(t), the unnecessary component X_i(t) and noise component X_n(t) are eliminated from the reception signal X(t) by the matrix R⁻¹. Only the desired wave component X_d(t) can be extracted. That is, only the reception signal of the transmission wave 4 arriving from the direction θ₁can be extracted.

[0095] Similarly, to extract the reception signals of the transmission waves 4 arriving from the directions θ₂to θ_Dthe weight vectors W are obtained on the basis of the steering vectors A_iof the unnecessary waves to obtain desired wave components W*X(t), respectively.

[0096] A delay time change measuring algorithm for causing the delay time change measuring unit 45 to measure a change in delay time of the transmission wave 4 will be described below.

[0097] The desired wave extracting unit 43 extracts 2K+1 IF (f₀±Δf×K) components from the spectra of the reception signals frequency-converted. by the FFT processor 41. The desired wave extracting unit 43 multiplies the respective frequency spectrum components with a weight to extract each frequency spectrum component of the desired wave.

[0098] A matrix Y₁represented by equation (21) is generated on the basis of this frequency spectrum components y_−K, y_−K+1, . . . y₀, y₁, . . . , y_K(complex numbers) of the desired wave:

[0099] Y₁=[y_−K, y_−K+1, . . . , y₀, y₁, . . . , y_K]  (21)

[0100] A signal having a frequency spectrum of the transmission wave 4 at predetermined time is defined as a reference signal. A matrix V represented by equation (22) is generated on the basis of IF (f₀±Δf×K) components v_−K, v_−k+1, . . . , v₀, v₁, . . . , v_K(complex numbers) of the reference signal:

[0101] V=[v_−K, v_−K+1, . . . , v₀, v₁, . . . , v_K]  (22)

[0102] As described above, since the A/D converters 31 to 3M sample the reception signals at intervals of integer multiples of the periods of the transmission waves 4, the delay amounts of the reception signals with reference to the predetermined time can be known by comparing the sampled reception signals with the reference signal.

[0103] FIGS. 4A and 4B explain reception signals sampled by the A/D converters 31 to 3M. FIG. 4A shows a transmission wave 4awhen radiated from the transmitting device 1, and FIG. 4B shows a transmission wave 4bwhen received by the linear antenna array 10. The abscissa in FIG. 4A or 4B represents time.

[0104] Assume that the period of the transmission wave 4ais defined as ΔT, and that the frequency spectrum of the reference signal is that of the transmission wave 4aat time T₀. In this case, the frequency spectrum of the reference signal is equivalent to that of the transmission wave 4aat time T₀+nΔT (n is a natural number).

[0105] When the reception signal sampled at time T1 is compared with the reference signal to obtain the delay time of the reception signal of the transmission wave 4aat time T₀, the reception signal sampled at time T₁+nΔT is compared with the reference signal to obtain the delay time of the reception signal of the transmission wave 4aat time T₀+nΔT, thereby obtaining the change in delay time during nΔT.

[0106] Each frequency component of the reception signal represented by equation (21) is divided by each frequency component of the reference signal represented by equation (22) to normalize the matrix Y₁. The normalized matrix is given as Y:

[0107] Y=[y_−k/v_−k, y_−k−1, . . . , y₀/v₀, . . . , y₁/v₁, . . . , y_K/v_K]  (23)

[0108] The respective frequency components of the normalized reception signal are linear coupling results of the D arrival signals F′₁, F′₂, . . . , F′_D(complex numbers) and the noise frequency components n′_−K, n′−_K−1, . . . , n′_K(complex numbers) as in equation (1). That is,

[0109] Y=A′F′+N′  (24)

[0110] In this case, the array manifold A′ is expressed by:

[0111] A′=[a′(τ₁), a′(τ₂), . . . , a′(τ_D)]  (25)

[0112] and the steering vector a′(τ) is expressed by:

[0113] α′(τ)=[e^{−j2π(f0−Δf×K)τ}, . . . , e^−j2πf0τ, . . . , e^{−j2π(f0+Δf×K)τ}]^T  (26)

[0114] This equation represents the reception power of each frequency component by the linear antenna array 10 with the delay time τ.

[0115] The delay time from the predetermined time T₀is estimated using a technique similar to the arrival direction measuring algorithm.

[0116] More specifically, the covariance matrix T_2K+1, of equation (24) is obtained as follows:

[0117] T_2K+1={overscore (YY  (27)

[0118] Moving space averaging is performed for equation (27). If the covariance matrix T_2K+1undergone moving space averaging is given as T, the following equation holds:

[0119] T=A′P′A′^H+σ²I  (28)

[0120] In this case, the following equation holds:

[0121] P′={overscore (F′F′  (29)

[0122] When the eigenvector for the eigenvalue of the noise component of the covariance matrix T is given as E′_N, the evaluation function for estimating the delay time can be expressed as follows:

[0123] P_MU(τ)=1/(a^H(τ)E′_NE′_N^Ha(τ))  (30)

[0124] so that a peak is searched using this evaluation function to estimate a delay time with respect to the predetermined time T₀.

[0125] If the delay times from the predetermined time T₀which are estimated on the basis of the reception signals sampled by the A/D converters 31 to 3M at different timings, e.g., (i−1)th and ith times (i≧2) are defined as τ_i−1and τ_i, respectively, a change Δτ in delay time during the sampling interval nΔT is obtained as follows:

[0126] Δτ=τ_i−τ_i−1  (31)

[0127] The operation of the multiple propagation wave parameter measuring apparatus shown in FIG. 1 will be described with reference to FIG. 5. FIG. 5 shows the operation flow of the signal processor 40 shown in FIG. 2.

[0128] As shown in FIG. 1, when the transmission waves 4 are radiated from the transmitting device 1, the receiving side directly receives some of the transmission waves 4. The reflecting members 3 reflect some of the transmission waves 4, and the receiving side receives the reflected transmission waves 4 as well.

[0129] The M antenna elements 11 to 1M constituting the linear antenna array 10 receive the transmission waves 4 passing through various routes, as shown in FIG. 2.

[0130] The reception signals output from the antenna elements 11 to 1M are converted by the receivers 21 to 2M into IFs which can be digitally processed, respectively.

[0131] The reception signals converted into the IFs are simultaneously sampled by the A/D converters 31 to 3M at the intervals of the integer multiples of the periods of the transmission waves 4. The sampled signals are converted into digital signals, which are then input to the signal processor 40.

[0132] The M reception signals output from the A/D converters 31 to 3M to the signal processor 40 are input to the FFT processor 41. An input reception signal may also be referred to as an “ith input reception signal” (steps S1 and S2).

[0133] The M reception signals input to the FFT processor 41 are FFT-processed and frequency-analyzed (step S3).

[0134] The reception signals frequency-analyzed by the FFT processor 41 are output to the arrival direction measuring unit 42 and subjected to the moving space averaging and MUSIC method, thereby measuring the arrival directions θ of the transmission waves 4 (step S4).

[0135] The reception signals frequency-analyzed by the FFT processor 41 are also output to the desired wave extracting unit 43. The desired wave extracting unit 43 extracts the reception signal of the transmission wave 4 arriving from each direction θ and outputs each extracted reception signal to the delay time change measuring unit 45 (step S5).

[0136] The reception signals output from the desired wave extracting unit 43 are subjected to the moving space averaging and MUSIC method in the delay time change measuring unit 45 to measure changes Δτ in delay times of the transmission waves 4 arriving from the respective directions θ (step S6).

[0137] The flow then returns to step S2 to read reception signals output from the A/D converters 31 to 3M again. The input reception signals. are (i+1)th input reception signals (step S7).

[0138] As described above, steps S2 to S7 are repeated.

[0139] FIG. 6 shows the operation flow of the arrival direction measuring unit 42 shown in FIG. 2. The operation of the arrival direction measuring unit 42 will be described in detail with reference to FIG. 6.

[0140] The arrival direction measuring unit 42 extracts IF f₀components of the M reception signals frequency-analyzed by the FFT processor 41 to generate the matrix X represented by equation (2) (step S41). The arrival direction measuring unit 42 generates the covariance matrix S_MMof the matrix X, as indicated by equation (7) (step S42).

[0141] The arrival direction measuring unit 42 performs moving space averaging for the covariance matrix S., to suppress the correlation of the arrival signals. At this time, the number m of rows and columns of the sub-array matrix S_mmobtained by dividing the covariance matrix S_MMis preset to be larger than the number D of the arriving transmission waves 4.

[0142] The resultant covariance matrix is defined as S (step S43).

[0143] The arrival direction measuring unit 42 decomposes the covariance matrix S into eigenvalues to obtain m eigenvalues represented by equation (12) (step S44). The arrival direction measuring unit 42 then calculates an eigenvector E_Nfor the minimum one of the m eigenvalues (step S45).

[0144] The arrival direction measuring unit 42 calculates the evaluation function P_MU(θ) represented by equation (16) using a predetermined angular interval to find a peak, thereby estimating the arrival direction θ of the transmission wave 4 (step S46). Data representing the arrival direction θ estimated in step S46 is output to the desired wave extracting unit 43 shown in FIG. 2.

[0145] FIG. 7 shows the operation flow of the desired wave extracting unit 43 shown in FIG. 2. The operation of the desired wave extracting unit 43 will be described in detail with reference to FIG. 7.

[0146] Assume that the arrival direction measuring unit 42 estimates the number of arrival waves as D.

[0147] The desired wave extracting unit 43 defines the transmission wave 4 arriving from one direction θ₁as a desired wave and the transmission waves 4 arriving from the directions other than the direction θ₁as unnecessary waves (step S51). The desired wave extracting unit 43 generates the matrix R represented by equation (19) using the steering vectors A_ias the responses of the linear antenna relay 10 to the directions except for the direction θ1 and the noise components a σ₂of the receivers 21 to 2M (step S52).

[0148] The desired wave extracting unit 43 calculates the weight vector W represented by equation (20) on the basis of the steering vector A_das the response of the linear antenna relay 10 to the direction θ₁and the inverse matrix of the matrix R (step S53).

[0149] The FFT processor 41 analyzes the frequencies of the M reception signals. The desired wave extracting unit 43 receives the 2K+1 IF (f₀±Δf×K) components of the respective frequency-analyzed reception signals (step S54). The desired wave extracting unit 43 multiplies each IF component with the weight vector W to extract the IF (f₀±Δf×K) component of the desired wave (step S55).

[0150] The desired wave extracting unit 43 repeats steps S52 to S55 D times to extract the IF components of the transmission waves 4 arriving from the respective directions. The desired wave extracting unit 43 outputs them to the delay time change measuring unit 45 (steps S52 to S57).

[0151] FIG. 8 shows the operation flow of the delay time change measuring unit 45. The operation of the delay time change measuring unit 45 will be described in detail with reference to FIG. 8.

[0152] The delay time change measuring unit 45 generates the matrix Y₁represented by equation (21) using the IF f₀±Δf×K component of the desired wave output from the desired wave extracting unit 43 (steps S61 and S62).

[0153] The delay time change measuring unit 45 divides the frequency components of the matrix Y₁by the frequency components of the reference signal stored in the storage unit 44 shown in FIG. 2 to normalize the matrix Y₁, thereby obtaining the matrix Y represented by equation (23) (step S63).

[0154] As described above, since the reference signal is a signal having the frequency spectrum at desired time T₀of the transmission wave 4, each component of the normalized matrix Y serves as a signal representing a delay state with respect to the predetermined time T₀.

[0155] The delay time change measuring unit 45 then generates the covariance matrix T_2K+1is of the matrix Y, as represented by equation (27) (step S64).

[0156] The delay time change measuring unit 45 performs moving space averaging for the covariance matrix T_2K+1to suppress the correlation between the frequencies. The resultant covariance. matrix is defined as T (step S65).

[0157] The delay time change measuring unit 45 estimates the delay time of the desired wave with respect to the predetermined time T₀by the MUSIC method.

[0158] More specifically, the delay time change measuring unit 45 decomposes the covariance matrix T into eigenvalues (step 66) to calculate the eigenvector E′_Nfor the minimum eigenvalue (step S67). The delay time change measuring unit 45 calculates the evaluation function P_MU(τ) represented by equation (30) at the predetermined time interval to find a peak, thereby estimating the delay time of the desired wave with respect to the predetermined time T₀.

[0159] The calculated delay time is the delay time of the ith reception signal input in steps S1 and S2 of FIG. 5 and defiled as τ_i(step S68).

[0160] If no delay time τ_i−1of the (i−1)th reception signal input in steps S1 and S2 of FIG. 5 is present, that is, if i=1, then the flow advances to step S71 (NO in step S69).

[0161] The delay time change measuring unit 45 repeats steps S62 to S69 D times to estimate the delay times τ₁of the transmission signals arriving from the respective directions with respect to the predetermined time T₀(steps S62 to S69 and steps S71 and S72).

[0162] If the delay time τ_i−1is present in step S69, that is, if i≧2, then the flow advances to step S70 (YES in step S69). The delay time change measuring unit 45 calculates the change Δτ in delay time of the desired wave at the sampling interval nΔT of the A/D converters 31 to 3M shown in FIG. 2 on the basis of the difference between the delay times τ₁and τ_i−1.

[0163] The delay time change measuring unit 45 repeats steps S62 to S70 D times to estimate the changes Δτ in delay times of the transmission waves 4 arriving from the respective directions and output them (steps S62 to S72).

[0164] As described above, the signal processor 40 measures the arrival directions θ of the arriving transmission waves 4 and extracts the reception signal having only the desired wave from the all the transmission waves arriving from the measured directions θ. The signal processor 40 then measures the change in delay time of each desired wave.

[0165] Assume that search for direction is performed at the interval of ΔΘ in the angular range of 0 to Θ, and that search for delay time is performed at the time interval of Δt in the time range of 0 to t. The conventional method requires calculations (Θ)/ΔΘ)×(t/Δt) times. To the contrary, the multiple propagation wave parameter measuring method of the present invention requires search for direction only once and search for delay time only D times. Therefore, (Θ/ΔΘ)×1+(t/Δt)×D is required to greatly shorten the processing time.

[0166] The signal processor 40 normalizes only the reception signal having the desired wave output from the desired wave extracting unit 43. However, the reception signals frequency-analyzed by the FFT processor may be normalized, and the reception signal having the desired wave may be extracted from the normalized frequency-analyzed reception signals.

[0167] In this case, the storage unit 44 in FIG. 2 is connected to the input of the desired wave extracting unit 43, and the desired wave extracting unit 43 has the function of normalizing the reception signals frequency-analyzed by the FFT processor 41. This process is performed between steps S54 and S55.

[0168] The moving distance Δd of the transmitting device 1 shown in FIG. 1 can be obtained by the change Δτ of the delay time of the transmission wave 4, which is measured by the receiving device 2 shown in FIG. 1. In this case, only the wave directly reaching the receiving device 2 from the transmitting device 1 is considered.

[0169] The change Δτ in delay time, which is measured by the receiving device 2, is the change Δτ in delay time at the sampling interval nΔT of the A/D converters 31 to 3M shown in FIG. 2. Therefore, the velocity c of light is multiplied with the change Δτ in delay time to obtain the moving distance Δd of the transmitting device 1 during the time nΔT.

[0170] The functions of the FFT processor 41, arrival direction measuring unit 42, desired wave extracting unit 43, and delay time change measuring unit 45 in the signal processor 40 shown in FIG. 2 can be realized by controlling a computer with a program. This program is called a multiple propagation wave parameter measuring program.

[0171] FIG. 9 shows the arrangement of a signal processor in a multiple propagation wave parameter measuring apparatus according to the second embodiment. The remaining components, i.e., a transmitting device 1, a linear antenna array 10 constituting a receiving device 2, receivers 21 to 2M, and A/D converters 31 to 3M are identical to those in the multiple propagation wave parameter measuring apparatus shown in FIG. 1, and a detailed description thereof will be omitted.

[0172] As shown in FIG. 9, a signal processor 50 comprises a computer. This computer comprises a CPU (Central Processing Unit) 51, a storage device 52 such as a hard disk, a first interface 53, and a second interface 54. The CPU 51, the storage device 52, and the first and second interfaces 53 and 54 are connected to each other via a bus 55.

[0173] The program for controlling the CPU 51 is stored in a machine-readable recording medium 60 and provided. The recording medium 60 may be, e.g., a flexible disk, CD-ROM, or memory card.

[0174] The program recorded on the recording medium is installed in the storage device 52 via the first interface 53 of the computer. The CPU 51 operates in accordance with the program stored in the storage device 52 and realizes the same functions as those of the FFT processor 41, arrival direction measuring unit 42, desired wave extracting unit 43, and delay time change measuring unit 45 shown in FIG. 2. The CPU 51 measures arrival directions θ and changes Δτ in delay times of transmission waves 4 on the basis of reception signals input from the A/D converters 31 to 3M.

[0175] As has been described above, according to the present invention, the arrival directions of transmission waves are measured, and the reception signal having a desired wave arriving from one of the measured arrival directions is extracted to measure the delay time and change on the basis of the extracted signal. In detecting a change in delay time of one desired wave, changes in delay times in all the directions need not be measured. The measuring time for measuring the “arrival direction” and “change in delay time” can be shortened.

[0176] In detecting a change in delay time of each desired wave, the changes in delay times in all the directions need not be measured, thereby shortening the measuring time.

[0177] According to the present invention, the reception signals are sampled at the intervals of integer multiples of the periods of the transmission waves, and the sampled reception signals are normalized by the frequency data of the transmission waves at predetermined time. The delay times are then obtained at the predetermined time. In addition, the difference between the delay times is calculated using the two reception signals sampled at different timings, and then changes in delay times are calculated. For this reason, even if a sync signal is not sent from the transmitting side to the receiving side via a cable, the change in delay time can be measured with high precision. Therefore, the change in delay time can be measured by moving at least one of the transmitting and receiving sides.