JPH1048308A - Apparatus and method for detecting object position - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable even positions of a plurality of objects to be detected by outputting predetermined signals to the objects and receiving the reflected signals with a plurality of sensors to calculate the impulse responses or the like of the respective sensors from the output signals and received signals. SOLUTION: Signals are radiated from signal radiating means 1a such as a speaker, and antenna to a space, and the radiated signals are directly or from an object 6 or the like to reach a plurality of sensors 2a, 2b or the like. Signals of the respective sensors and signals of the means 1a are processed by a method of the use of a suited filter or the like to calculate the impulse response by the respective sensors. Then, the weight of an imagined position in each imagined position set to be grid points for example in a reserching space is figured out according to a predetermined system on the basis of the respective impulse responses to obtain the spatial distribution. A peak is detected from the distribution and the position is outputted as an object position. Thus, the positions of a plurality of objects can be presumed. Also, the use of a plurality of the signal outputting means improves the accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object position detecting device and an object position detecting method for automatically detecting an object position.

[0002]

2. Description of the Related Art Heretofore, detection of an object at a doorway of a house or indoors has been performed by processing image information input by a video camera or detecting changes in radiated radio waves or light by a sensor. However, these methods cannot detect when the object is in the shadow or outside the field of view of the camera. Therefore, Japanese Patent Application Laid-Open No. 7-146366 discloses a method for detecting an object having a shadow by utilizing the diffraction effect of a sound wave. This method radiates a sound wave, obtains an acoustic transfer characteristic from the reverberation sound, and obtains a position of an object based on a difference between the transfer characteristic when there is no object and when there is a certain object. At this time, one sound source and a plurality of sensors or a plurality of sound sources outputting the same signal and one sensor measure an impulse response, which is a time-domain expression of acoustic transfer characteristics, and detect an object position. ing.

[0003]

However, in the above disclosed technology, since information of one object closest to the measurement point is extracted, only one object can be detected, and it is desired to detect a plurality of objects simultaneously. There was a problem that the case could not be met.

The present invention has been made in view of such a problem, and an object of the present invention is to provide an object position detecting apparatus and an object position detecting apparatus capable of accurately detecting the position even when there are a plurality of objects. It is to provide a detection method.

[0005]

In order to achieve the above object, an object position detecting apparatus according to a first aspect of the present invention provides a signal output means for generating a predetermined signal and outputting the signal to an arbitrary object toward a space. A signal input unit for individually inputting a signal reflected from the object by a plurality of sensors; a signal output from the signal output unit; and a signal input to the plurality of sensors. Impulse response calculation means for obtaining an impulse response for each, a virtual position is determined at an arbitrary point, and the propagation time until the output signal output to the space reaches the signal input means as reflected at the virtual position is determined. Determine, weight calculation means for calculating the weight of the virtual position from the components of each impulse response calculated based on this propagation time, and calculating the weight while shifting the virtual position, Weight is and a object position estimating means for estimating that the position of the object virtual position larger than a prescribed threshold value.

In the object position detecting device according to a second aspect of the present invention, in the first aspect, the signal output means includes a plurality of signal emission means for outputting a predetermined signal to a space toward an arbitrary object; Signal generating means for generating signals individually radiated through the plurality of signal emitting means, wherein the impulse response measuring means measures impulse responses corresponding to the plurality of signal emitting means, The position estimating means estimates a position of the object based on at least a part of the plurality of impulse responses.

According to a third aspect of the present invention, in the first or second aspect, the impulse response measuring means includes an initial state estimation mode selecting means for estimating an impulse response in an initial state; An initial impulse response subtracting means for obtaining an impulse response by subtracting the impulse response of the state.

The object position detecting method according to a fourth aspect of the present invention provides a signal output step of generating a predetermined signal and outputting the signal to an arbitrary object toward a space, and a plurality of sensors for outputting a signal reflected from the object. A signal inputting step of individually inputting, a signal output in the signal outputting step, and an impulse response calculating step of obtaining an impulse response for each sensor based on the signals input to the plurality of sensors; A virtual position is determined at one point, a propagation time until the output signal output to the space reaches the signal input position as reflected at the virtual position is determined, and each impulse response calculated based on this propagation time is calculated. A weight calculation step of calculating the weight of the virtual position from the components of the above, and the weight is calculated while shifting the virtual position, and the virtual position at which the weight is equal to or greater than a predetermined threshold value Comprising the object position estimating step of estimating that the position of the object.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, an outline of the present embodiment will be described. In the present embodiment, the weight of a virtual position arbitrarily set in the search range is calculated from a plurality of impulse response components obtained from each of the input signals of the plurality of sensors, and a position where this weight has a large value is determined as a reflection object. By using the position, even if there are a plurality of objects, the position can be estimated. The weight of the virtual position is calculated from a component of the impulse response calculated based on a propagation time from the signal source to the sensor through the virtual position to reach the sensor from the signal source for measuring the impulse response.

An embodiment of the present invention will be described below in detail with reference to the drawings. First, an outline of a first embodiment will be described with reference to FIG. In FIG. 1, an object position detection device according to the present embodiment includes a signal output unit 1 that generates a signal for estimating an impulse response and outputs the signal to a space such as an indoor space.
A signal input unit 2 for inputting a signal output to space and reflected back from an object by a plurality of sensors; an impulse response calculation unit 3 for estimating an impulse response from an output signal and an input signal; An object position estimating unit 4 for determining an object position from the impulse response.

In this configuration, a signal is radiated from the signal output unit 1 to the space, and the radiated signal reaches the sensor of the signal input unit 2 directly or by being reflected by a surrounding object. As a result, signals reflecting the surrounding environment are captured by a plurality of sensors installed at different positions. The impulse response calculator 3 calculates an impulse response between each sensor and the signal output from the signal output unit 1 for each sensor. The object position estimating unit 4 including the weight calculating means obtains a weight for each virtual position set in the search space based on the plurality of obtained impulse responses. The position where the value of the weight shows a large value is obtained as the sound source position.

According to the above-described method, the number of detectable objects is not limited to one, and a plurality of object positions can be estimated. Here, a signal radiating unit such as a speaker or an antenna that actually outputs a signal to a space and a sensor that inputs a signal are placed in a room, for example, as shown in FIG. 2 (the signal radiating unit 1a, Sensors 2a, 2b). Reference numeral 6 denotes an object to be detected. The signal output from the signal radiation means 1a is, for example, white noise, impulse, swept sine wave band noise, or the like, and in the case of a sound wave, it may be in an audible band or a higher frequency band. The signal output to the space may be a sound wave or a radio wave, but will be described as a sound wave for convenience. Further, it is desirable that the radiation characteristic of the signal radiation means 1a be omnidirectional or a characteristic having a loose directivity that is uniform over a range where an object to be detected exists.

The calculation of the impulse response in the impulse response calculation unit 3 is performed, for example, according to the literature (Transactions of the Institute of Electronics, Information and Communication Engineers D-II vol. J77-D-II No. 6pp.1).
037-1047 (1994.6)), may be performed using an adaptive filter, or may be performed using a cross spectrum method based on a fast Fourier transform. Is desirably a method using an adaptive filter. In addition, an impulse may be output to a space as a signal, and a waveform input from a sensor may be used as an impulse response as it is. FIG. 3 shows a configuration of the impulse response calculation unit 3 using an adaptive filter. The impulse response calculation unit 3 is composed of adaptive filters of several sensors (1 to M). In this case, the signal input to the signal input unit 2 is read out for each sensor channel, and each adaptive filter 3a (1 to M) is read. The signal output from the signal output unit 1 is input as a desired signal of each adaptive filter 3a.

From the above two inputs, each adaptive filter 3a
, A new impulse response is calculated for each sample of the signal, and at the same time, a signal in which the output signal component in the input signal is canceled is calculated. The adaptive filter may be a well-known Least Mean Square (LMS) method,
ive Least Square (RLS). Details of the adaptive filter are described in detail in a document (Heiken: Introduction to Adaptive Filters). Here, FIG. 4 illustrating measurement of an impulse response by a Normalized LMS (NLMS) which is one of LMS adaptive filters is described. , NLMS
4 shows a configuration for estimating an impulse response of a transmission system using an adaptive filter. In FIG. 4, this configuration is
A convolution operation unit 11 for calculating a convolution x * h of an FIR filter h representing an impulse response and an input x;
The apparatus includes a filter update unit 12 that updates a filter coefficient based on y−x * h, a power calculation unit 13 that calculates the power p of an input signal, an adder 14, and a multiplier 15. x is an input signal and y is a desired signal.

The impulse response h is updated by the following equation: h _j = h _j-1 −a * e * x / 2p (1) As a result of the update, the impulse response which is a transfer characteristic between the desired signal and the input signal is obtained. Is obtained, and at the same time, an error signal e is output. The error signal e is a signal obtained by canceling the desired signal component y from the input signal x. h _j is the impulse response after j updates, a is the step size,
It is experimentally determined in the range of 0 <a <1.0. For example, 0.1 or the like is used. Also, when the power of the input signal is smaller than the noise, the estimation error is prevented from increasing due to the noise. As shown in FIG.
In addition, when the value of the power p of the input signal is smaller than a predetermined threshold value, adaptive control processing for preventing the filter from being updated is generally performed. Here, adaptive control information as to whether or not to update the adaptive filter is also output outside the adaptive filter operation unit.

The value of the power of the input signal is obtained by calculating the average power from the 128 samples from the present time of the desired response input to the adaptive filter main body to, for example, 128 points in the past, and applying the adaptive control information of the adaptation stop / continue to one sample. Output each time. The adaptive filter main body performs an adaptive operation based on the adaptive control information. The above threshold value is set to, for example, a value 20 dB lower than the average value of the radiation output signal as the desired signal.

Next, the processing of the object position estimating unit 4 performed based on a plurality of impulse responses will be described. For example, it is assumed that a signal output unit, a sensor, and an object are arranged as shown in FIG.

At this time, an output signal is output from the signal emitting means 1a, propagates at a propagation speed c, is reflected by the object 6, and
The propagation time to reach the th sensor is expressed as Ti = (r _so + r _oi ) / c. The component of the impulse response is considered to represent the magnitude of the reflected wave component for each time delay, and a time delay component corresponding to this propagation time is extracted from each impulse response. Since the component of the normal impulse response is obtained at each sampling period, the time delay closest to the propagation time Ti of the reflection signal from the object at the virtual position may be rounded off, or the sampling period may be obtained by linear interpolation as in the following equation. The value of the impulse response at a point with a time delay that is not an integral multiple of may be determined.

Wi = (n + 1−Ti) * hi (n−1) + (Ti−n) * hi (n) (2) where Ts is a sampling period, n is an integer obtained by truncating Ti / Ts, hi (K) is the value at the point of the time delay k of the impulse response of the i-th sensor. Moreover, the impulse Wi = focused on the energy of the response (n + 1-Ti) * | hi (n-1) | 2 + (Ti-n) * | hi (n) | may be ^two (3). In consideration of attenuation due to propagation, Wi = (n + 1- Ti) * | hi (n-1) / (r so + r oi) | 2 + (Ti-n) * | hi (n) / (r so + r _oi ) | ² (4).

After the above components corresponding to the propagation time are obtained for each impulse response, the sum ΣWi of them is taken as the weight of the virtual position, and this is calculated over the entire search range by changing the position of the virtual position. . At this time, a product may be used instead of the sum of the weights. For the virtual position, for example, as shown in FIG. 7, a plurality of grid points are set in a search range, weights are calculated for all the set grid points, and a spatial distribution of weights is obtained. Finally, a peak is detected from the obtained distribution of weights, and the position is output as an object position.

Next, the flow of the above processing will be described with reference to FIG. First, as an initial setting, the signal radiation means 1
The position of a, the position of the sensor 2a, the search range, the step size of the search, and the like are set (step S1).

Next, the signal output processing is sufficiently long, for example, 6
The random noise sequence for 0 seconds is stored in the memory,
The data is read out, DA converted, and output from the speaker. When the data is completed, the data is returned to the beginning of the data and read out, and this is repeated until the entire processing is completed (step S2).

Next, the signal input processing is the AD conversion of the signal of the sensor in the case of a sound wave. For example, the sampling frequency is set to 40 kHz and stored in a ring buffer for one second, for example, and the entire processing is completed. Continue until. This AD conversion is performed for all the sensors (step S3).

The signal fetching process of the signal input unit 2 and the process of generating an output signal and outputting to a space in the signal output unit 1 are operated in parallel independently of other processes.

Next, the input signal is converted into one block length, for example, 10 blocks.
Data is read from the ring buffer for 24 points. Reading is performed for all sensors (step S4). Next, the output signal is read out from the memory storing the output signal for one block length, for example, for 1024 points. Reading is performed from the beginning of the data. When reading is performed in the next step, reading is performed from the next data after the last data of the currently read data, and when reading to the end of the data is performed, returning to the beginning of the data is repeated ( Step S5).

Next, the impulse response is calculated and updated by the adaptive filter from the read input data and output data of one block length. The calculation is performed for all adaptive filters (step S6).

Next, a virtual position is set in the search space,
The weight of the virtual position is calculated by, for example, Expression (3). The virtual position is, for example, a grid point set in the search range, as shown in FIG. 7, calculates weights for all the set grid points, and obtains a spatial distribution of the weights (step S7).

Next, a peak is obtained from the distribution of the weights obtained in step S7, and a peak position at which the weight of the peak position is equal to or greater than the determined threshold is output as an object position (step S7).
8). Here, the threshold value for peak detection can be set, for example, to a level 3 dB higher than the average value of the weight distribution over the search range, but it is better to experimentally determine it according to the situation in which it is performed.

Steps S4 to S8 are repeated until the processing is completed. It should be noted that the signal to be used in the present invention may be any signal for which an impulse response can be estimated.
Not only sound waves but also electric waves may be used. Also, the target space is
Not only in the air, but also in the water or the ground.

According to the above-described first embodiment, it is possible to estimate the positions of a plurality of objects, and to obtain a signal waveform obtained by canceling the output signal components from the signals of the plurality of sensors. The signal waveform with this output signal component canceled is
It can also be used as a signal from a microphone array for voice recording.

Next, a second embodiment of the present invention will be described. The present embodiment outputs signals to a space from a plurality of signal emitting means, and obtains an impulse response corresponding to each signal emitting means, thereby substantially estimating an object position estimation accuracy equivalent to increasing the number of sensors. Is what makes it possible to get That is, since the number of impulse responses corresponding to the number of signal radiating means is determined for one sensor, M * N where M is the number of sensors and N is the number of signal radiating means.
M * N impulse responses are obtained using the number of adaptive filters, so that the number of additions when obtaining the weight of the virtual position increases, and the estimation accuracy is improved.

The overall configuration of this embodiment is the same as that of the above-described first embodiment, except that the configuration of the signal output section and the impulse response calculation section is changed so as to support a plurality of signal outputs.

FIG. 9 is a diagram showing a configuration of a signal output unit according to the second embodiment. The signal output unit 1 includes a signal generation unit 1a that generates an output signal, and a plurality of signal emission units 1b (1 to N).
Each output signal is output to the space from the signal radiating unit 1b and is also output to the impulse response calculating unit 3 at the same time. As in the first embodiment, the signal generation unit 1a stores a sufficiently long signal sequence, for example, for 60 seconds in a memory, reads it out, sends it to the signal radiating unit 1b, and when the data ends, puts it at the beginning of the data. It returns to read and repeats this, and continues until the whole processing is completed. At this time, a different random signal sequence having no correlation is prepared for each signal radiating unit 1b, and another signal sequence is transmitted for each signal radiating unit 1b.

The impulse response calculator 3 has a configuration as shown in FIG. 10 so as to calculate an impulse response between each of the plurality of output signals and each sensor input signal.
In FIG. 10, the number of adaptive filters for calculating the impulse response is equal to the number of sensors * the number of signal radiating sections.
Input signal of M from each sensor 1 and signal output unit 1 to N
The impulse response is measured using each of the above signals as an input.

Signals from a plurality of signal output means are superimposed and input to the sensor. Of the signal components of the sensor, components having no correlation with the signals from the signal output units 1 to N, which are the desired inputs of the adaptive filter, This is not a problem because it is treated as noise and its effect decreases over the observation time.

In the case of the second embodiment, the processing procedure is different from that of the first embodiment in that the number of impulse response calculations increases from the number of sensors to the number of sensors * the number of signal output means.
Since the number of impulse responses used in the object position estimating process only increases from M to M * N and there is no essential difference in the processing procedure, it will not be described again.

Hereinafter, a third embodiment of the present invention will be described.
In the second embodiment, a case where an uncorrelated random signal is output to space for estimating an impulse response is taken as an example. In this case, since the output signal acts as noise, the estimation accuracy of the impulse response may decrease. However, the convergence speed of the adaptive filter may be reduced. In this embodiment, in order to avoid these problems, a description will be given of a case where signals that do not overlap with each other are output and a case where a signal including components in different frequency domains is output between the signal output units 1. .

When signals are output at different times from each other, data contents stored as signals in the signal generator are as shown in FIG. 11, for example. FIG. 11 shows the contents of signal data corresponding to the signal emitting units 1 to N,
At the same time, there is a signal only for the data of one signal output means,
Others are silent. Also, in consideration of reverberation, a short silence section, for example, 500 ms, indicated by To in FIG. 11, for example, may be inserted before and after each signaled section in order to take a time during which all signal radiating sections 1b are silent. The data in the signaled section may be a pulse sequence or a random signal.

When the impulse response measurement is performed using the above-described signals, the adaptive filter of the impulse response measurement unit 3 determines the filter coefficient according to the level of the radiation output signal, which is the desired signal, as described in the first embodiment. Control of the suspension / continuation of the update is performed to prevent adaptation due to noise.

The above processing flow is the same as the processing flow described in the second embodiment. Next, a case will be described in which the signal output means outputs signals composed of components in different frequency domains. In this case, the data stored in the signal generation unit 1a has frequency characteristics as shown in FIG. 12, for example. FIG. 12 shows the signal radiating section 1.
It shows frequency characteristics of signal data corresponding to b (1 to N), each of which has a comb-shaped band. In each comb-shaped frequency characteristic, a frequency band having a large power is prevented from overlapping with each other in the signal radiation portions. The signal waveform having such a comb-shaped frequency characteristic is
For example, it can be generated by superimposing a plurality of sine waves of different frequencies.

In the case of the signal output from the signal output unit 1 shown in FIG. 12, the frequency band having the power ranges from fa1 to fb shown in FIG.
1, sine waves having frequencies in these bands are represented by sin (2πk (fbi-fai) / (N−1)). , (1 <= i <5, 0 <k <= N−1). Here, i is the number of bands having power, N is the number of sine waves, for example, 10, k is the number of sine waves, and if N is increased, the band will be covered more densely.

When calculating the impulse response using the signals of the different frequency bands, the impulse response calculator 3
Has a configuration as shown in FIG.
Here, reference numeral 20 denotes a comb filter, and reference numeral 21 denotes an adaptive filter main body, which has a configuration shown in FIGS. 4 and 5, for example.
FIG. 13 shows a configuration in which the signal passed through the comb filter 20 is supplied to each adaptive filter main body 21, and the characteristics of the comb filter 20 are such that the output to be input to the adaptive filter main body 21 to which the comb filter 20 is connected is shown. Make it match the frequency characteristics of the signal. The processing flow is the same as that of the second embodiment except that a step of passing the comb filter is added to the step of the adaptive filter operation (FIG. 8, step 6).

Next, a fourth embodiment of the present invention will be described. In the fourth embodiment, the intensity of the signal output from the signal output unit to the space is changed according to the position of the object. As described above, in this embodiment, a signal such as a sound wave or a radio wave is output to a space for impulse response measurement. When the position is to be detected by a human or when a human is present in the search range, the output signal needs to be unpleasant and harmless to the human. For example, when outputting a sound wave, it is harsh in the audible band, and even if a frequency band outside the audible band is used, it may harm the human body if the intensity is high. This effect is considered to be greater as the distance between the signal radiating means, for example, the speaker and the person is shorter. Therefore, when an object is detected, the distance between the detected object and the signal emitting means is obtained, and the intensity of the output signal is changed based on this distance.

For the above purpose, in the fourth embodiment, the second
The configuration of the signal output unit of the embodiment is configured as shown in FIG. 14, and the information on the object position from the object position estimating unit is input.

In FIG. 14, 1d is a signal radiating section, 1c
Is an output signal generation unit. When the output signal generation unit 1c generates signals of the number N channels of the signal emission unit 1d, the intensity is changed based on the object position estimated by the object position estimation unit 4. ing. The signal radiating section 1d is 1
It goes without saying that the same configuration can be applied to the configuration of the first embodiment in which the number N of the signal radiation units 1d is set to one.

Next, the processing of the signal output section including the output signal generation section 1c will be described with reference to FIG. First, an applicable maximum distance Rmax is set in an initial setting step to set an applicable range when an output signal is changed according to a distance from an object (step S1). This is a value provided to change the output signal only when an object within a range shorter than the applicable maximum distance is detected, and is to prevent a response to a distant object having low detection accuracy.

Next, data of one block, for example, 1024 points is read from the memory storing the output signal data (step S2). Next, the information of the object position detection from the object position estimating unit 4 is checked (step S3). If the object is detected, the process proceeds to the next step S4. If the object is not detected, the process proceeds from the signal emitting unit 1d. The signal of this block is output to space (step S7), and the process returns to step S2.

Next, in step S4, a distance R between the object position and the signal output means is obtained. When there are a plurality of signal output means, the smaller value is R. Next, the above R is compared with the applicable maximum distance Rmax. If R <Rmax, the process proceeds to the next step S6, and if not, the process proceeds to step S7 (step S5).

In the next step S6, first, in step S2
The amplitude of the read signal data is made R / Rmax times.
Next, signal data is sent to the signal radiating section 1d to output a signal to the space (step S7), the process returns to step S2, and the same is repeated thereafter.

Next, a fifth embodiment of the present invention will be described. In the fifth embodiment, the intensity of the signal output from the signal output unit to the space is changed according to the position of the object. As described above, in the present embodiment, a signal such as a sound wave or a radio wave is output to a space in order to measure an impulse response. When the position detection target is a human, a reflected sound component from a fixed object may be strongly included. When it is desired to detect only a moving object excluding these fixed objects, a simple method is to subtract a time-invariant weight component from the distribution of weights estimated by the object position estimating unit, thereby obtaining It is possible to detect only moving objects that change. On the other hand, the position of a moving object can be detected by subtracting a time-invariant component from among the components of the impulse response first and estimating the object position using the remaining impulse responses. In this embodiment, a moving object is estimated using the difference between the impulse responses.

FIG. 16 is a diagram showing the configuration of the fifth embodiment, in which the measurement mode control unit 22 for instructing to measure only the component of the impulse response that is temporally invariant when there is no moving object is provided by the first mode. This is a configuration added to FIG. 1, which is the basic configuration of the fifth embodiment to the fifth embodiment. Here, whether to measure an impulse response as a base component to be subtracted or to output a result obtained by subtracting the base component as an impulse response is referred to as a measurement mode.

The base impulse response measurement may be instructed by a user using a switch, a computer keyboard, a mouse, or the like, or automatically performed for a predetermined time, for example, 5 seconds immediately after the start of the object position detection processing. Alternatively, the base impulse response measurement mode may be set and normal impulse response estimation may be performed after 5 seconds have elapsed. When measuring automatically for a certain period of time as in the latter case, the measurement mode control unit 22
For example, it can be realized by a configuration as shown in FIG.

In FIG. 17, reference numeral 23 denotes a time counter,
Reference numeral 24 denotes a measurement mode generation unit, which measures time from the start of processing by a time counter 23, and
Thus, the base impulse response measurement mode is output before the lapse of a predetermined time, and the normal impulse response measurement mode is output after the lapse of a predetermined time. The time can be easily measured by a computer system, for example, a UNIX system function.

Accordingly, each adaptive filter of the impulse response calculator 3 is configured as shown in FIG. In FIG. 18, reference numeral 25 denotes an adaptive filter main body for measuring an impulse response, which has, for example, the configuration shown in FIGS.
26 is a base impulse response subtraction unit. The impulse response measured by the adaptive filter main body 25 is sent to the base impulse response subtraction unit 26, and the processing result is sent to the object position estimation unit 4. The base impulse response subtractor 26 stores the measured impulse response as the base impulse response when the measurement mode is the base measurement, and subtracts the base impulse response from the transmitted impulse response when the measurement mode is the normal measurement.

The processing flow of the present embodiment including the above-described base impulse response subtraction processing will be described with reference to FIG. First, calculation of the impulse response by the adaptive filter from the initial setting (step S1) (step S1)
Up to 6) is the same as in the first embodiment. Next, in step S7, it is checked whether the measurement mode is the base measurement or the normal measurement. If the measurement mode is the base measurement, the impulse response is stored as the base impulse response (step S8), and the process returns to step S4. If it is a normal measurement, the next step S
Go to 9. In the next step S9, the base impulse response is subtracted from the measured impulse response. Next, the estimation processing of the object position is performed in the same manner as in the first embodiment (steps S10 to S11).

The above steps S4 to S11 are repeated until the processing is completed. In the object position estimating apparatus described in the fifth embodiment, when estimating the impulse response,
By removing the reflection component from the fixed object, the position of a plurality of moving objects can be estimated.

The above-described impulse response calculation step, weight calculation step, and object position estimation step are stored as computer programs in a storage medium such as a hard disk, a floppy disk, or a CD-ROM. It can be installed and executed on a simple computer.

[0058]

According to the present invention, it is possible to detect a plurality of object positions based on impulse responses at a plurality of measurement points. Since the amount of calculation required for space search processing required for object position estimation is extremely small, it is suitable for real-time processing. In addition, by performing processing using a plurality of signal output units, the number of impulse responses used for object estimation can be substantially increased, and an object position can be accurately detected with a small number of sensors. .

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of an object position detection device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an example of the arrangement of signal radiating means such as a speaker and an antenna and a sensor for inputting a signal;

FIG. 3 is a diagram illustrating a configuration of an impulse response calculation unit.

FIG. 4 is a diagram showing a configuration of an adaptive filter.

FIG. 5 is a diagram illustrating a configuration of an adaptive control unit.

FIG. 6 is an explanatory diagram of a propagation time.

FIG. 7 is a diagram for describing setting of a virtual position.

FIG. 8 is a flowchart illustrating a flow of a process according to the first embodiment.

FIG. 9 is a diagram illustrating a configuration of a signal output unit of an object position detection device according to a second embodiment of the present invention.

FIG. 10 is a diagram showing a configuration of an impulse response calculation unit.

FIG. 11 is a diagram showing the contents of signal data when signals are output such that output times are different from each other.

FIG. 12 is a diagram illustrating frequency characteristics of each signal radiation unit.

FIG. 13 is a diagram illustrating a configuration of an adaptive filter when a plurality of output signals having different frequency characteristics are used.

FIG. 14 is a diagram illustrating a configuration of a signal output unit of an object position detection device according to a fourth embodiment of the present invention.

FIG. 15 is a flowchart illustrating a flow of processing of a signal output unit according to a fourth embodiment.

FIG. 16 is a diagram illustrating a configuration of an object position detection device according to a fifth embodiment of the present invention.

FIG. 17 is a diagram illustrating a configuration of a measurement mode control unit.

FIG. 18 is a diagram illustrating a configuration of an adaptive filter including a base subtraction unit.

FIG. 19 is a flowchart illustrating the flow of a process according to the fifth embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Signal output part, 2 ... Signal input part, 3 ... Impulse response calculation part, 4 ... Object position estimation part.

Claims (4)

[Claims] 1. A signal output means for generating a predetermined signal and outputting the signal to a space toward an arbitrary object; a signal input means for individually inputting a signal reflected from the object by a plurality of sensors; An impulse response calculating means for obtaining an impulse response for each sensor based on a signal output from the output means and a signal input to the plurality of sensors; determining a virtual position at an arbitrary point and outputting to the space; A propagation time until the output signal obtained reaches the signal input means as reflected at the virtual position is obtained, and a weight of the virtual position is calculated from each impulse response component calculated based on the propagation time. Weight calculating means; calculating the weight while shifting the virtual position; and estimating the virtual position where the weight is equal to or greater than a predetermined threshold as the position of the object. Object position detecting device characterized by comprising an estimating means. 2. The signal output means includes: a plurality of signal emitting means for outputting a predetermined signal to a space toward an arbitrary object;
Signal generating means for generating signals individually radiated through the plurality of signal emitting means, wherein the impulse response measuring means measures impulse responses corresponding to the plurality of signal emitting means, 2. The object detection device according to claim 1, wherein the position estimating unit estimates the position of the object based on at least a part of the plurality of impulse responses. 3. The impulse response measuring means includes an initial state estimation mode selecting means for estimating an impulse response in an initial state, and an initial impulse response subtracting means for obtaining an impulse response by subtracting the impulse response in the initial state. 3. The object position detecting device according to claim 1, wherein 4. A signal output step of generating a predetermined signal and outputting the signal to a space toward an arbitrary object; a signal input step of individually inputting signals reflected from the object by a plurality of sensors; An impulse response calculation step of obtaining an impulse response for each sensor based on the signal output in the output step and the signals input to the plurality of sensors, determining a virtual position at any one point, and outputting the virtual position to the space A propagation time required for the output signal obtained to reach the signal input position assuming that the output signal is reflected at the virtual position, and a weight for calculating the weight of the virtual position from each impulse response component calculated based on the propagation time A calculating step, calculating the weight while shifting the virtual position, and estimating an object position in which a virtual position where the weight is equal to or greater than a predetermined threshold is estimated to be the position of the object Object position detecting method characterized by comprising the estimated step.