JP2019215275A - Target detection device, and target detection method - Google Patents

Abstract

To provide a target detection device and target detection method that can upgrade responsiveness with respect to a target.SOLUTION: A target detection device 1 comprises: a generation unit 32a; a calculation unit; and an estimation unit 323b. The generation unit is configured to generate instant data including a relative speed toward an own vehicle of a target as to each of a plurality of reflection points where a transmitted radio wave is reflected upon a target. The calculation unit is configured to calculate a ground speed on the basis of the relative speed of the instant data generated by the generation unit and a speed of the own vehicle. The estimation unit is configured to make origins in a ground vector toward the own vehicle corresponding to each of a plurality of instant data uniform on the basis of the ground speed calculated by the calculation unit, and estimate a movement direction of the target from an intersection of a perpendicular line with respect to the ground vector.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a target detection device and a target detection method.

  Conventionally, for example, a target detection device that transmits a radio wave around a vehicle and detects instantaneous data including a relative speed of a direction of the target toward the own vehicle based on a reflected wave of the transmitted radio wave reflected by the target is known. It is known (for example, see Patent Document 1).

JP 2000-193745 A

  However, the relative speed detected by the target detection device described above is only the speed component in the direction to the own vehicle, and in order to obtain the speed component in the actual moving direction of the target, it takes several scans from the initial detection. Instantaneous data is required. Therefore, the conventional target detection device has a problem that the responsiveness to the target cannot be improved because the actual moving direction of the target is known after several scans.

  The present invention has been made in view of the above, and an object of the present invention is to provide a target detection device and a target detection method capable of improving responsiveness to a target.

  In order to solve the above-described problem and achieve the object, a target detection device according to the present invention includes a generation unit, a calculation unit, and an estimation unit. The generation unit generates, for each of a plurality of reflection points at which the transmitted radio wave is reflected by the target, instantaneous data including a relative speed of a direction of the target toward the own vehicle. The calculation unit calculates a ground speed based on the relative speed of the instantaneous data generated by the generation unit and the speed of the host vehicle. The estimating unit, based on the ground speed calculated by the calculating unit, aligns the starting point in the ground vector of the direction to the own vehicle corresponding to each of the plurality of instantaneous data, from the intersection of the perpendicular to the ground vector. Estimate the moving direction of the target.

  According to the present invention, responsiveness to a target can be improved.

FIG. 1A is a diagram illustrating an outline of a target detection method according to the embodiment. FIG. 1B is a diagram illustrating an outline of a target detection method according to the embodiment. FIG. 1C is a diagram illustrating an outline of the target detection method according to the embodiment. FIG. 2 is a block diagram illustrating a configuration of the radar device according to the embodiment. FIG. 3 is an explanatory diagram of processing from the pre-processing of the signal processing unit to the peak extraction processing in the signal processing unit. FIG. 4A is an explanatory diagram of the angle estimation process. FIG. 4B is an explanatory diagram (part 1) of the pairing process. FIG. 4C is an explanatory diagram (part 2) of the pairing process. FIG. 5 is a diagram illustrating a prediction process based on a ground speed. FIG. 6 is a functional block diagram of the estimation unit. FIG. 7 is a diagram illustrating an update process of the buffer update unit. FIG. 8 is a diagram illustrating a process of calculating a ground speed by a ground speed calculation unit. FIG. 9A is a diagram showing a method of creating a movement vector when there are three intersections. FIG. 9B is a diagram showing a method of creating a movement vector when there are six intersections. FIG. 10 is a diagram illustrating a weighting process based on a movement vector. FIG. 11 is a flowchart illustrating a processing procedure of a process executed by the target detection device according to the embodiment. FIG. 12 is a flowchart illustrating a processing procedure of an estimation process performed by the estimation unit according to the embodiment. FIG. 13 is a diagram illustrating the verification processing of the movement vector.

  Hereinafter, embodiments of a target detection device and a target detection method disclosed in the present application will be described in detail with reference to the accompanying drawings. The present invention is not limited by the embodiment. In the following, a case where the target detecting device is an FM-CW (Frequency Modulated Continuous Wave) method will be described as an example. However, the target detecting device may use another method such as an FCM (Fast-Chirp Modulation) method. The system may be used.

  First, an outline of a target detection method according to the embodiment will be described with reference to FIGS. 1A to 1C. 1A to 1C are diagrams illustrating an outline of a target detection method according to the embodiment. FIG. 1A shows a host vehicle MC on which the target detection device 1 according to the embodiment is mounted, and another vehicle LC that moves with a predetermined movement vector V. The movement vector V is a vector including the actual movement direction of the other vehicle LC and the ground speed in the movement direction. The ground speed is a speed obtained by removing the speed component of the own vehicle MC from the relative speed of the own vehicle MC and the other vehicle LC. In addition, the own vehicle MC moves with a predetermined movement vector MV (hereinafter, own vehicle vector MV).

  As shown in FIG. 1A, the target detection device 1 is mounted in, for example, a front grill of the own vehicle MC, and detects a target (for example, another vehicle LC or the like) existing in the traveling direction of the own vehicle MC. Note that the target detection device 1 may be mounted on another location such as a windshield, a rear grill, and left and right side portions (for example, left and right door mirrors).

  Further, as shown in FIG. 1A, the target detection device 1 generates instantaneous data 100 for each of a plurality of reflection points at which radio waves transmitted around the own vehicle MC are reflected by other vehicles LC. The instantaneous data 100 includes, for example, the relative speed (vector RV) of the direction to the host vehicle MC.

  Here, a conventional target detection device will be described. In the conventional target detection device, the instantaneous data detected in one scan is only the relative speed in the direction toward the host vehicle (for example, the vector RV shown in FIG. 1A), and the above-described movement vector cannot be obtained. That is, in one scan, the actual moving direction of the target cannot be obtained, and in order to obtain the actual moving direction of the target, it is necessary to obtain the actual moving direction of the target from a change in instantaneous data for several scans from the first detection. Therefore, the conventional target detection device has a problem that the responsiveness to the target cannot be improved because the actual moving direction of the target is known after several scans.

  Therefore, in the target detecting method according to the embodiment, the actual moving direction of the target can be detected in one scan. Specifically, as shown in FIG. 1B, in the target detection method according to the embodiment, first, the speed of the host vehicle MC is calculated from the relative speed in the vector RV (hereinafter, relative vector RV) of the plurality of generated instantaneous data 100. A vector GV (hereinafter, ground vector GV) indicating the ground speed in the direction toward the own vehicle MC excluding the components is calculated. Specifically, the ground vector GV is a vector obtained by projecting the movement vector V onto a straight line passing through the own vehicle MC and the instantaneous data 100. The speed component of the host vehicle MC is a vector obtained by projecting the host vehicle vector MV onto the above-described straight line. This point will be described later with reference to FIG.

  Subsequently, as shown in FIG. 1C, in the target detection method according to the embodiment, based on the ground speed in the plurality of instantaneous data 100, the ground vector GV of the direction to the own vehicle MC corresponding to each of the plurality of instantaneous data 100. Are aligned, and the moving direction of the other vehicle LC as the target is estimated from the intersection CP of the perpendicular VL extending from the end point of the ground vector GV.

  The example illustrated in FIG. 1C illustrates a two-dimensional plane in which the starting points of the ground vectors GV corresponding to the two instantaneous data 100 are aligned. In FIG. 1C, the vertical axis indicates the ground speed (VY) in the left-right direction with respect to the host vehicle MC, and the horizontal axis indicates the ground speed (VX) in the front-rear direction with respect to the host vehicle MC.

  As shown in FIG. 1C, in the target detection method according to the embodiment, first, a perpendicular VL passing through the end points of two ground vectors GV whose starting points are aligned is drawn, and an intersection CP of the two perpendiculars VL is obtained. In the target detection method according to the embodiment, when there is one intersection CP, the starting point of the ground vector GV is set as the starting point of the movement vector V, and the intersection CP is set as the end point of the movement vector V. That is, the direction from the starting point of the ground vector GV to the intersection CP indicates the actual moving direction of the target, and the distance between the starting point of the ground vector GV and the intersection CP indicates the ground speed in the moving direction of the target.

  As described above, according to the target detection method according to the embodiment, the moving direction of the target can be detected in one scan based on the plurality of instantaneous data 100. Responsiveness can be improved.

  Further, in the target detection method according to the embodiment, by estimating the movement vector V based on the ground speed of the instantaneous data 100, a change (speed, position, and the like) of the own vehicle MC is considered as compared with the relative speed. Since there is no need, the estimation accuracy of the movement vector V can be further improved.

  In the target detection method according to the embodiment, if only the moving direction of the target is necessary, it is not always necessary to obtain the distance between the starting point of the ground vector GV, which is the ground speed in the moving direction, and the intersection CP. Absent. That is, in the target detection method according to the embodiment, at least the moving direction of the target is estimated.

  In the target detection method according to the embodiment, when the number of intersections CP is three or more, the moving direction of the target is estimated by obtaining the inner center of a triangle formed by the intersections CP. Will be described later.

  Next, the configuration of the target detection device 1 according to the embodiment will be described in detail with reference to FIG. FIG. 2 is a block diagram illustrating a configuration of the target detection device 1 according to the embodiment. In FIG. 2, only the components necessary for describing the features of the present embodiment are represented by functional blocks, and descriptions of general components are omitted.

  In other words, each component illustrated in FIG. 2 is a functional concept and does not necessarily need to be physically configured as illustrated. For example, the specific form of distribution / integration of each functional block is not limited to the illustrated one, and all or a part of the functional blocks may be functionally or physically distributed in arbitrary units according to various loads and usage conditions. -It is possible to integrate and configure.

  As illustrated in FIG. 2, the target detection device 1 includes a transmission unit 10, a reception unit 20, and a processing unit 30. The target detection device 1 is connected to a vehicle control device 2 that controls the behavior of the host vehicle MC.

  The vehicle control device 2 performs vehicle control such as a PCS (Pre-crash Safety System) and an AEB (Advanced Emergency Braking System) based on the detection result of the target by the target detection device 1.

  The transmission unit 10 includes a signal generation unit 11, an oscillator 12, and a transmission antenna 13. The signal generation unit 11 generates a modulation signal for transmitting a millimeter wave frequency-modulated by a triangular wave under the control of a transmission / reception control unit 31 described later. The oscillator 12 generates a transmission signal based on the modulation signal generated by the signal generation unit 11, and outputs the transmission signal to the transmission antenna 13. Note that, as shown in FIG. 2, the transmission signal generated by the oscillator 12 is also distributed to a mixer 22, which will be described later.

  The transmission antenna 13 converts a transmission signal from the oscillator 12 into a transmission wave, and outputs the transmission wave to the outside of the vehicle MC. The transmission wave output by the transmission antenna 13 is a continuous wave frequency-modulated by a triangular wave. A transmission wave transmitted from the transmission antenna 13 to the outside of the own vehicle MC, for example, forward, is reflected by a target such as another vehicle LC and becomes a reflected wave.

  The receiving unit 20 includes a plurality of receiving antennas 21 forming an array antenna, a plurality of mixers 22, and a plurality of A / D converters 23. The mixer 22 and the A / D converter 23 are provided for each reception antenna 21.

  Each receiving antenna 21 receives a reflected wave from the target as a received wave, converts the received wave into a received signal, and outputs the signal to the mixer 22. Although the number of receiving antennas 21 shown in FIG. 2 is four, it may be three or less or five or more.

  The reception signal output from the reception antenna 21 is input to the mixer 22 after being amplified by an amplifier (not shown) (not shown). The mixer 22 mixes a part of the distributed transmission signal and a part of the reception signal input from the reception antenna 21, removes unnecessary signal components, generates a beat signal, and outputs the beat signal to the A / D conversion unit 23. .

  The beat signal has a beat frequency that is the difference between the frequency of the transmission signal (hereinafter, referred to as “transmission frequency”) and the frequency of the reception signal (hereinafter, referred to as “reception frequency”). The beat signal generated by the mixer 22 is converted into a digital signal by the A / D converter 23 after the timing is adjusted between the receiving antennas by a synchronization unit (not shown), and then output to the processing unit 30.

  The processing unit 30 includes a transmission / reception control unit 31, a signal processing unit 32, and a storage unit 33. The signal processing unit 32 includes a generation unit 32a and a filter processing unit 32b.

  The storage unit 33 stores the history data 33a. The history data 33a is information including the history of the target data and the history of the instantaneous data 100 in a series of signal processing executed by the signal processing unit 32. The history data 33a also includes information such as the ground speed of the direction to the own vehicle MC of the instantaneous data 100 and an angle corresponding to the direction to the own vehicle MC. Such a point will be described later.

  The processing unit 30 is, for example, a microcomputer including a CPU (Central Processing Unit), a ROM (Read Only Memory) and a RAM (Random Access Memory) corresponding to the storage unit 33, a register, and other input / output ports. The entire target detection device 1 is controlled.

  When the CPU of the microcomputer reads out and executes the program stored in the ROM, the microcomputer functions as the transmission / reception control unit 31 and the signal processing unit 32. Note that the transmission / reception control unit 31 and the signal processing unit 32 may be entirely configured by hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

  The transmission / reception control unit 31 controls the transmission unit 10 including the signal generation unit 11 and the reception unit 20. The signal processing unit 32 periodically executes a series of signal processing. Subsequently, each component of the signal processing unit 32 will be described.

  The generation unit 32a generates the instantaneous data 100. Specifically, the generation unit 32a generates the instantaneous data 100 by performing a frequency analysis process, a peak extraction process, and an instantaneous data generation process.

  In the frequency analysis processing, a fast Fourier transform (FFT: Fast Fourier Transform) processing (hereinafter, referred to as “FFT processing”) is performed on the beat signal input from each A / D conversion unit 23. The result of the FFT processing is a frequency spectrum of the beat signal, and a power value (signal level) for each frequency of the beat signal (for each frequency bin set at a frequency interval corresponding to the frequency resolution).

  In the peak extraction processing, a peak frequency that becomes a peak in the result of the FFT processing by the frequency analysis processing is extracted. In the peak extraction processing, a peak frequency is extracted for each of an “UP section” and a “DN section” of a beat signal described later.

  In the instantaneous data generation process, an angle estimation process for calculating an arrival angle and a power value of a reflected wave corresponding to each of the peak frequencies extracted in the peak extraction process is executed. At the time of execution of the angle estimation process, the arrival angle is an angle at which the target is estimated to be present, and thus may be referred to as an “estimated angle” below.

  In the instantaneous data generation process, a pairing process of determining a correct combination of the peak frequencies of the “UP section” and the “DN section” based on the calculation result of the calculated estimated angle and the power value is executed.

  In the instantaneous data generation processing, the distance of each target to the own vehicle MC and the relative speed of the direction to the own vehicle MC are calculated from the determined combination result. In the instantaneous data processing, the calculated estimated angle, distance, and relative speed of each target are output to the filter processing unit 32b as the instantaneous data 100 for the latest cycle (latest scan), and the history data 33a in the storage unit 33 is output. To be stored.

  In order to make the description easy to understand, FIGS. 3 to 4C show the flow of processing from the pre-stage processing of the signal processing unit 32 to the processing in the signal processing unit 32. FIG. 3 is an explanatory diagram of a process from a pre-stage process of the signal processing unit 32 to a peak extraction process in the generation unit 32a.

  FIG. 4A is an explanatory diagram of the angle estimation process. FIGS. 4B and 4C are explanatory diagrams (part 1) and (part 2) of the pairing process. FIG. 3 is divided into three regions by two thick downward white arrows. Hereinafter, each of these areas is described as an upper row, a middle row, and a lower row.

  As shown in the upper part of FIG. 3, the transmission signal fs (t) is transmitted from the transmission antenna 13 as a transmission wave, is reflected by a target, arrives as a reflected wave, and is received by the reception antenna 21 at the reception signal fr (t). ).

  At this time, as shown in the upper part of FIG. 3, the reception signal fr (t) is delayed from the transmission signal fs (t) by a time difference T according to the distance between the host vehicle MC and the target. Due to the time difference T and the Doppler effect based on the relative speeds of the host vehicle MC and the target, the beat signal has a frequency fup in an “UP section” where the frequency increases and a frequency fdn in a “DN section” where the frequency decreases. Are obtained as a repeated signal (see the middle part of FIG. 3).

  The lower part of FIG. 3 schematically shows the result of the FFT processing of the beat signal in the frequency analysis processing for each of the “UP section” side and the “DN section” side.

  As shown in the lower part of FIG. 3, after the FFT processing, waveforms in the respective frequency regions on the “UP section” side and the “DN section” side are obtained. In the peak extraction processing, a peak frequency that becomes a peak in such a waveform is extracted.

  For example, in the case of the example shown in the lower part of FIG. 3, a peak extraction threshold is used. On the “UP section” side, peaks Pu1 to Pu3 are determined as peaks, respectively, and peak frequencies fu1 to fu3 are extracted.

  On the “DN section” side, the peaks Pd1 to Pd3 are determined as peaks by the same peak extraction threshold, and the peak frequencies fd1 to fd3 are respectively extracted.

  Here, the frequency component of each peak frequency extracted by the peak extraction processing may include a mixture of reflected waves from a plurality of targets. Therefore, in the instantaneous data generation process, an angle estimation process for calculating the azimuth for each of the peak frequencies is performed, and the presence of a target corresponding to each peak frequency is analyzed.

  The azimuth calculation in the instantaneous data generation processing can be performed using a known arrival direction estimation technique such as ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques).

  FIG. 4A schematically shows the azimuth calculation result of the instantaneous data generation process. In the instantaneous data generation processing, from the peaks Pu1 to Pu3 of the azimuth calculation result, the estimated angles of the targets (each reflection point) respectively corresponding to these peaks Pu1 to Pu3 are calculated. The magnitude of each of the peaks Pu1 to Pu3 is a power value. In the instantaneous data generation process, as shown in FIG. 4B, the angle estimation process is performed on each of the “UP section” side and the “DN section” side.

  Then, in the instantaneous data generation process, a pairing process of combining peaks having close estimated angles and power values in the direction calculation result is performed. Further, from the combination result, in the instantaneous data generation processing, the distance of each target (each reflection point) corresponding to each combination of peaks and the relative speed of the direction to the own vehicle MC are calculated.

  The distance can be calculated based on the relationship of “distance∝ (fup + fdn)”. The relative speed can be calculated based on the relationship of “speed∝ (fup−fdn)”. As a result, as shown in FIG. 4C, a pairing processing result indicating the instantaneous data 100 of the estimated angle, the distance, and the relative speed of each reflection point RP with respect to the host vehicle MC is obtained.

  Returning to FIG. 2, the filter processing unit 32b will be described. As illustrated in FIG. 2, the filter processing unit 32b includes a prediction unit 321b, an assignment unit 322b, an estimation unit 323b, a weighting unit 324b, a resampling unit 325b, and a target data generation unit 326b.

  The filter processing unit 32b generates target data corresponding to the instantaneous data 100 by applying a particle filter that allocates a predetermined number of particle data to the instantaneous data 100 generated by the generation unit 32a.

The prediction unit 321b performs a process of predicting sample points (particle data) in the particle filter. Probability Specifically, the prediction unit 321b includes the latest period time t, when the distribution X _t of the particle data at time t, based on the distribution state X _t-1 time t-1 of the previous cycle N pieces of particle data are arranged (sampled) based on the density function. That is, in the prediction process, the prediction unit 321b distributes the particle data from the particle data at the time t−1 to an area where the instantaneous data 100 is likely to appear at the time t.

  Further, the prediction unit 321b generates prediction data corresponding to the current target data based on the previous target data. Specifically, the prediction unit 321b generates prediction data based on the previous moving direction of the target data and the relative speed in the moving direction.

  Further, the prediction unit 321b distributes the particle data in a predetermined distribution state for the instantaneous data 100 corresponding to the new target because there is no particle data in the previous cycle.

  Note that the prediction unit 321b is not limited to the case where the prediction data is generated based on the relative speed in the moving direction. For example, the prediction unit 321b may generate the prediction data based on the ground speed in the moving direction. Here, the prediction process using the ground speed will be described with reference to FIG.

  FIG. 5 is a diagram illustrating a prediction process based on a ground speed. The method of calculating the ground speed will be described later. In FIG. 5, when the target detection device 1 (t) at time t turns rightward and moves to the target detection device 1 (t + 1) at time t + 1, the target data 50 (t) at time t A prediction process for predicting the prediction data 60 (t + 1) at the time t + 1 will be described. In other words, the prediction process is a process of predicting the target data itself and a process of performing a coordinate system conversion accompanying the movement of the vehicle MC. In the following, a description will be given by taking as an example a procedure of first performing prediction processing of the target data itself and then performing coordinate system conversion. Note that there is no problem even if the coordinate system conversion is performed first, and then the prediction processing is performed.

  As shown in FIG. 5, the prediction unit 321b first moves the target data 50 (t) by the movement vector V (ground speed) to generate temporary data 50 (t + 1) before the coordinate system conversion. In other words, the destination of the target in the coordinate system based on the position and direction of the vehicle MC at the time t at the previous measurement is predicted as the temporary data 50 (t + 1).

  Based on the state of the host vehicle MC at time t, the position and direction of the target detection device 1 (t + 1) at time t + 1 which is the reference direction after the coordinate system conversion is predicted. More specifically, the prediction unit 321b predicts the position of the target detection device 1 (t + 1) at time t + 1 based on the traveling speed and the turning radius of the vehicle MC. More specifically, the prediction unit 321b predicts a position change (Δx and Δy) on the XY plane from time t to time t + 1 and a change in direction, that is, a rotation angle θ. That is, at this time, the translational movement amount and the rotational movement amount of the vehicle MC are obtained.

  Then, the prediction unit 321b determines the predicted position of the prediction data 60 (t + 1) based on the position change (Δx and Δy) and the rotation angle θ. That is, the coordinate system conversion is performed on the temporary data 50 (t + 1) based on the translation amount and the rotation amount. Specifically, the prediction unit 321b first performs translational conversion on the temporary data 50 (t + 1). Specifically, the movement is performed based on the position change (Δy and Δx) described above. Next, rotational movement conversion is performed. Specifically, the temporary data 50 (t + 1) after the translation conversion is rotated by the rotation angle θ. If the translational conversion is performed first, the rotation center in the rotational movement conversion, that is, the origin, and the actual rotation center, that is, the position of the target detection device 1 (t + 1) are convenient. That is, the prediction unit 321b performs the prediction when the position and the direction of the target detection device 1 at the time t + 1 are set to the origin and the reference direction based on the amount of change of the target detection device 1 from the time t to the time t + 1. The position of the data 60 (t + 1) is generated. This makes it possible to predict a position change from the target data 50 (t) to the prediction data 60 (t + 1) when the target detection devices 1 (t) and 1 (t + 1) are aligned with the origin on the XY plane. In the above example, the translation conversion and the rotation conversion are separately applied, but this is not a limitation. For example, in the case of the translational transformation and the rotational transformation, a combination of the operations can be described as one linear transformation matrix. The rotation amount is described using the rotation angle θ as a variable, but is not limited thereto. There is no problem in using a method of describing using a direction vector, a complex number, a quaternion, or the like. In addition, there are various coordinate system conversion methods, and various methods can be selected as required.

  Returning to FIG. 2, the assignment unit 322b will be described. The allocating unit 322b performs a process of allocating the instantaneous data 100 in the latest cycle to the latest particle data, which is the prediction result of the prediction unit 321b. Specifically, the allocating unit 322b allocates the particle data corresponding to the prediction data to the instantaneous data 100 existing within a predetermined allocation range corresponding to the prediction data generated by the prediction unit 321b.

  In addition, when there is the instantaneous data 100 that does not exist in the allocation range of any target data, the allocation unit 322b treats the instantaneous data 100 as a new target.

  The estimation unit 323b estimates the movement vector V of the target based on the plurality of instantaneous data 100 generated by the generation unit 32a. Specifically, the estimation unit 323b calculates the ground vector GV indicating the ground speed in the direction toward the own vehicle MC corresponding to each of the plurality of instantaneous data 100 based on the ground speed in the direction toward the own vehicle MC in the instantaneous data 100. The starting points are aligned, and the movement vector V of the target is estimated from the intersection CP of the perpendicular VL to the ground vector GV.

  Here, the processing content of the estimation unit 323b will be specifically described with reference to FIGS. FIG. 6 is a functional block diagram of the estimation unit 323b. As illustrated in FIG. 6, the estimation unit 323b includes an instantaneous data selection unit 323ba, a buffer update unit 323bb, a ground speed calculation unit 323bc (an example of a calculation unit), a buffering unit 323bd, and a perpendicular line creation unit 323be. An intersection creation unit 323bf, an inner center creation unit 323bg, and a movement vector creation unit 323bh are provided.

  The instantaneous data selection unit 323ba selects the instantaneous data 100 used for the process of estimating the movement vector V. Specifically, the instantaneous data selection unit 323ba determines the relative speed (or ground speed) of the direction to the own vehicle MC included in the instantaneous data 100, and the angle and the distance corresponding to the direction to the own vehicle MC. In the process of estimating the movement vector V, a process of excluding unnecessary instantaneous data 100 is performed.

  For example, the instantaneous data selection unit 323ba excludes the instantaneous data 100 in which the direction of the relative vector RV (or the ground vector GV) to the own vehicle MC is similar from the plurality of instantaneous data 100 from the estimation process of the movement vector V. . Specifically, the instantaneous data selection unit 323ba excludes one of the two instantaneous data 100 when the angle difference (for example, substantially zero) between the two instantaneous data 100 is less than a predetermined threshold.

  In such a case, the instantaneous data selection unit 323ba, for example, leaves the instantaneous data 100 having a shorter distance to the host vehicle MC and the instantaneous data 100 having a larger angle power value among the two instantaneous data 100. I do. This is because two instantaneous data 100 having substantially the same angle but different relative speeds (or ground speeds) cannot be obtained from the same target. Further, it is assumed that two instantaneous data 100 having substantially the same angle corresponding to the direction to the host vehicle MC and the same relative speed (or ground speed) are derived from reflection points at similar positions on the target. It is conceivable that in the process of estimating the movement vector V, if only one of them is used, the amount of information is sufficient, so that it is excluded.

  As described above, by removing at least one of the plurality of instantaneous data 100 having substantially the same angle and excluding the rest, the processing amount in the estimation process of the estimation unit 323b can be performed without lowering the estimation accuracy of the movement vector V. Can be prevented from increasing.

  Further, the instantaneous data selection unit 323ba estimates the instantaneous data 100 having a similar relative speed in the direction to the own vehicle MC and a similar distance to the own vehicle MC among the plurality of instantaneous data 100, by estimating the movement vector V. Exclude from

  Specifically, the instantaneous data selecting unit 323ba determines that the relative speed difference (for example, approximately zero) between the two instantaneous data 100 is less than a predetermined threshold value and the distance difference (for example, approximately zero) is less than a predetermined threshold value. One of the two instantaneous data 100 is excluded. In such a case, the instantaneous data selection unit 323ba leaves the instantaneous data 100 having the larger angle power value or the instantaneous data 100 having the angle closer to the target data. This is because the two instantaneous data 100 having substantially the same relative speed and the same distance do not exist in principle, and there is a high possibility that the instantaneous data 100 or the angle crack of a different target is present.

  Thus, it is possible to prevent a decrease in the estimation accuracy of the movement vector V due to the use of the instantaneous data 100 of another target, and to prevent an increase in the amount of processing in the estimation processing of the estimation unit 323b.

  Next, the buffer update unit 323bb updates the history data 33a stored in the storage unit 33. The update process by the buffer update unit 323bb is performed, for example, before adding the new instantaneous data 100 selected by the instantaneous data selection unit 323ba to the history data 33a, but may be performed at any timing as long as it is performed once in the measurement cycle. The execution timing may be appropriately set according to the relationship between the processing content and the buffer content. Specifically, when the new instantaneous data 100 is stored as the history data 33a, the buffer update unit 323bb uses the own vehicle of the past instantaneous data 100 in the history data 33a based on the change amount of the rotation angle of the own vehicle MC. The angle corresponding to the direction to the MC is corrected. Here, the update process of the buffer update unit 323bb will be described more specifically with reference to FIG.

  FIG. 7 is a diagram illustrating an update process of the buffer update unit 323bb. FIG. 7 shows the history data 33a before and after the update processing. Specifically, the history data 33a includes items such as “ID”, “relative speed”, “ground speed”, “angle”, and “storage progress count”.

  “ID” is identification information for identifying each instantaneous data 100. “Relative speed” indicates the relative speed of the direction to the host vehicle MC in the instantaneous data 100. The “ground speed” indicates the ground speed in the direction toward the host vehicle MC calculated by the ground speed calculation unit 323bc described below. “Angle” is the angle estimated in the angle estimation processing of the instantaneous data generation processing, and is a direction corresponding to the direction toward the host vehicle MC. Specifically, the “angle” indicates an angle in the left-right direction when the front direction of the vehicle MC is set to zero. The “storage progress count” indicates a cycle that has elapsed since the scan of the latest cycle. For example, the storage progress count “1” indicates that the current cycle is one cycle before the latest cycle, that is, the instantaneous data 100 obtained at time t−1.

  As illustrated in FIG. 7, the buffer update unit 323bb updates the “angle” and the “save progress count” of the history data 33a based on the amount of change in the rotation angle of the vehicle MC. The amount of change in the rotation angle of the host vehicle MC is a rotation angle in the horizontal direction accompanying the turning of the host vehicle MC. Specifically, the change amount of the rotation angle (referred to as θ) is calculated by the following equation, where Vself is the traveling speed of the own vehicle MC, crvR is the turning radius of the own vehicle MC, and Δt is the change amount of time. Is done. That is, it is calculated by θ = Vself × Δt / crvR. The direction of turning, that is, rightward or leftward, that is, positive or negative of θ is determined using information such as a steering angle.

  In the example illustrated in FIG. 7, it is assumed that the change amount (θ) of the rotation angle of the own vehicle MC is −3 degrees (for example, turning right forward). In such a case, the buffer update unit 323bb subtracts "-3 degrees" from the "angle" in all the past instantaneous data 100 included in the history data 33a. In terms of the amount of change, "3 times" is added. Further, the buffer updating unit 323bb increases the “save progress count” in all the past instantaneous data 100 included in the history data 33a by “1”. These past instantaneous data 100 are used when the number of new instantaneous data 100 is less than a predetermined number in the subsequent process of estimating the movement vector V.

  That is, the buffer update unit 323bb applies the coordinate system conversion accompanying the turning of the own vehicle MC to the past instantaneous data 100 as performed by the prediction unit 321b. That is, the stored past instantaneous data 100 also has the current direction in the coordinate system of the host vehicle MC, similarly to the present instantaneous data 100. Therefore, in the estimation processing of the movement vector V at the subsequent stage, the past instantaneous data 100 can be aligned with the same reference as the “angle” of the latest instantaneous data 100. Therefore, in the estimation process of the movement vector V, the past instantaneous data 100 can be used in the same manner as the current instantaneous data 100, and as a result, the estimation accuracy can be improved.

  FIG. 7 illustrates the case where the “angle” of the instantaneous data 100 is added. However, when the change amount of the rotation angle of the host vehicle MC is a positive value (for example, turning left forward), the instantaneous data is added. The value for the “angle” of 100 is subtracted. In the above, the method of correcting the angle of the history data 33a each time has been exemplified, but the present invention is not limited thereto. For example, there is a method in which only the change amount θn of the rotation angle of the host vehicle MC at each measurement timing n is recorded, and θn is continuously subtracted when using the history data 33a. In this case, the corrected angle θ for the history data 33a at the time tk is θ = θn + θn−1 +... + Θn−k. θn may be an integrated value, not the change amount of the rotation angle, that is, the own vehicle angle itself at each measurement timing n. In that case, θ = θn−θn−k. Instead of the speed and the turning radius, the direction of the own vehicle directly measured using an electronic compass or the like may be used. Although the angle has been exemplified, the direction vector may be used, or an operation using a complex number or a quaternion may be used. There are various methods for converting the coordinate system of the rotational movement, and the method can be appropriately selected according to the purpose.

  Also, as shown in FIG. 7, the buffer update unit 323bb deletes the instantaneous data 100 for which the “storage progress count” has become a predetermined number (“3” in FIG. 7) or more from the history data 33a. This is an operation for excluding the instantaneous data 100 whose reliability has been reduced due to the fact that the information becomes relatively old, and not using the instantaneous data 100 for estimation processing of the movement vector V at the subsequent stage. Can be improved. In the above description, deletion is performed when the elapsed count is exceeded, but the method is not limited to this. Conversely, there is a method of recording the life as a lifetime, subtracting the count in the update process, and excluding the count when the count reaches 0. There is also a method of recording only the count of the measurement time, not updating the history data 33a, and excluding the history data 33a past a predetermined count or more from the current measurement count. There are various methods, which can be appropriately selected according to the purpose.

  Next, the ground speed calculation unit 323bc calculates the ground speed by removing the speed component of the own vehicle MC from the relative speed of the direction to the own vehicle MC of the instant data 100 selected by the instant data selection unit 323ba. That is, the relative speed of the instantaneous data 100 includes an apparent speed component associated with the movement of the own vehicle MC and a speed component associated with the actual movement of the other vehicle, and the ground speed calculation unit 323bc calculates the latter The speed component accompanying the actual movement of another vehicle is extracted. That is, the ground speed can be said to be a speed component obtained by projecting the actual moving speed of another vehicle in the direction of the own vehicle in a stationary coordinate system, that is, a coordinate system based on the ground or the like. Here, the ground speed calculation process will be described with reference to FIG.

  FIG. 8 is a diagram illustrating a ground speed calculation process performed by the ground speed calculation unit 323bc. FIG. 8 shows an own moving vector MV indicating the actual moving direction of the own vehicle MC and the actual traveling speed in the moving direction. As shown in FIG. 8, the ground speed calculation unit 323bc first calculates the own vehicle vector MaV in the direction to the instantaneous data 100 based on the own vehicle vector MV. Specifically, the own vehicle vector MaV is a vector obtained by projecting the own vehicle vector MV on a straight line passing through the own vehicle MC and the instantaneous data 100. The host vehicle vector MaV is the speed component of the host vehicle MC in the relative vector RV.

  That is, the ground speed calculation unit 323bc calculates the ground vector GV indicating the ground speed in the direction toward the own vehicle MC by adding the relative vector RV of the instantaneous data 100 and the own vehicle vector MaV.

  Next, the buffering unit 323bd stores the new instantaneous data 100 selected by the instantaneous data selection unit 323ba in the history data 33a of the storage unit 33. The “storage progress count” of the new instantaneous data 100 in the history data 33a is appropriately set to an initial value such as “0”.

  Next, the perpendicular creating unit 323be creates a perpendicular VL to the ground vector GV indicating the ground speed in the direction toward the host vehicle MC of the instantaneous data 100 calculated by the ground speed calculating unit 323bc.

  Specifically, when the starting points of the ground vector GV corresponding to each of the plurality of instantaneous data 100 are aligned, the perpendicular creating unit 323be creates a perpendicular VL passing through the end point of the ground vector GV. For example, the perpendicular creating unit 323be calculates the origin of a plane represented by a ground speed component of the host vehicle MC in the left-right direction (VY axis) and a ground speed component of the host vehicle MC in the front-rear direction (VX axis). It is assumed that the starting points of the ground vector GV are aligned (see FIG. 1C).

  In such a case, the linear equation of the perpendicular VL is represented by VY = a × VX + b. Specifically, assuming that the angle α of the ground vector GV with respect to the VX axis and the ground vector GV is the ground speed Ta, the coefficient a is represented by a = tan (α + π / 2), The coefficient b is represented by b = Ta / sin (α). Note that when α = 0 or ± 2π, the coefficient a diverges. In the subsequent processing, exception processing is performed with Vx = V.

  When the number of the latest instantaneous data 100 is insufficient, the perpendicular creating unit 323be performs the perpendicular creating process using the past instantaneous data 100 of the history data 33a. That is, when the number of the latest instantaneous data 100 is less than the predetermined number, the perpendicular line creating unit 323be uses the history data 33a (an example of the past instantaneous data 100) stored in the storage unit 33 for estimating the movement vector V. .

  Accordingly, it is possible to suppress a decrease in the estimation accuracy of the movement vector V due to a shortage of the instantaneous data 100.

  Next, the intersection creating unit 323bf creates an intersection CP of the perpendicular VL created by the perpendicular creating unit 323be. Specifically, the intersection creating unit 323bf sets the two perpendiculars VL for finding the intersection CP to n1 and n2, and when the coefficient a of n1 is represented by a [n1] and the coefficient b is represented by b [n1], The coordinates VX are calculated by VX = (b [n2] −b [n1]) / (a [n1] −a [n2]), and the coordinates VY are obtained by substituting the above VX into VY = a × VX + b. Is calculated.

  Next, when there are three or more intersection points CP created by the intersection point creation unit 323bf, the inner center creation unit 323bg creates the triangle inner center IP (see FIG. 9A) formed by the three intersection points CP. Note that a method of creating the inner core IP by the inner core creating unit 323bg will be described later with reference to FIG. 9A.

  Next, the movement vector creation unit 323bh creates the movement vector V of the target based on the intersection CP created by the intersection creation unit 323bf. For example, when there is one intersection CP created by the intersection creation unit 323bf, the movement vector creation unit 323bh sets the starting point of the vector RV (the origin of the VY axis and the VX axis) as the starting point of the movement vector V, and Is estimated as the end point of the movement vector V.

  That is, based on the movement vector V created by the movement vector creation unit 323bh, the estimation unit 323b estimates that the direction from the starting point to the intersection CP in the ground vector GV is the actual movement direction of the target, and It is estimated that the distance between the starting point and the intersection CP in the GV is the ground speed in the moving direction of the target.

  In addition, when the number of intersection points CP created by the intersection point creation unit 323bf is three or more, the movement vector creation unit 323bh creates the movement vector V based on the inner center IP created by the inner center creation unit 323bg. Here, a method of creating the movement vector V when there are three or more intersection points CP will be described with reference to FIGS. 9A and 9B.

  FIG. 9A is a diagram illustrating a method of creating a movement vector V when there are three intersection points CP. FIG. 9B is a diagram illustrating a method of creating the movement vector V when there are six intersection points CP. First, with reference to FIG. 9A, a method of creating a movement vector V when there are three intersection points CP will be described.

  First, the inner center creating unit 323bg creates inner centers IP of the three intersections CP1, CP2, and CP3. Specifically, the inner center IP is the center of a triangular inscribed circle formed at the three intersections CP1, CP2, and CP3.

  More specifically, the inner center creating unit 323bg creates calculation vectors CV1, CV2, and CV3 starting from the origins on the planes of the VY axis and the VX axis and ending at the intersections CP1, CP2, and CP3.

  Then, the inner center creating unit 323bg creates an inner center IP based on the calculation vectors CV1, CV2, and CV3 and the opposite sides of the intersections CP1, CP2, and CP3. Specifically, when the opposite sides of the intersections CP1, CP2, and CP3 are represented as CP1a, CP2a, and CP3a, the inner-center creating unit 323bg obtains a vector (movement vector V in FIG. 9A) from the origin to the inner center IP by V = It is calculated by (CP1a × CV1 + CP2a × CV2 + CP3a × CV3) / (CP1a + CP2a + CP3a).

  Note that, when the variation of the three intersection points CP forming the triangle is equal to or more than a predetermined value, the inner center creating unit 323bg excludes the variation from the inner center IP calculation process. Specifically, when the unbiased variance of the three intersection points CP is equal to or more than a predetermined value, the inner center creating unit 323bg excludes the unbiased variance from the inner center IP calculation process.

  Then, when there is one inner core IP, the movement vector generator 323bh creates a vector from the origin to the inner core IP as the movement vector V. As described above, by calculating the inner center IP and creating the movement vector V, even if there are a plurality of intersections CP, it is possible to prevent the estimation accuracy of the movement vector V from being lowered.

  In addition, when it is clear that the error of any two calculation vectors CV1, CV2, and CV3 among the calculation vectors CV1, CV2, and CV3 is equal to or more than a predetermined value, the movement vector creation unit 323bh performs, for example, The remaining one calculation vector CV1, CV2, CV3 may be used as the movement vector V.

  Next, with reference to FIG. 9B, a method of creating the movement vector V in the case where the number of intersection points CP is six will be described. As shown in FIG. 9B, when there are six intersections CP3 to CP8, there are a plurality of triangles formed by the intersections CP3 to CP8, and thus there are also a plurality of inner centers IP1 to IP4. In the example shown in FIG. 9B, there is a combination in which a triangle is not formed by three intersections CP (for example, intersections CP3, CP5, and CP8), and such a combination of intersections CP is excluded from the creation of the inner center IP.

  As illustrated in FIG. 9B, the movement vector creation unit 323bh calculates, for example, an average point IPA that is an average value of the inner centers IP1 to IP4, and calculates a vector from the origin to the average point IPA as the movement vector V.

  As described above, when a plurality of inner core IPs are created, by creating the movement vector V based on the average point IPA, it is possible to minimize the estimation error of the movement vector V due to the variation of the inner core IP.

  Note that, when the variation of the plurality of inner cores IP is equal to or more than a predetermined value, the movement vector creation unit 323bh excludes the variation from the creation processing of the movement vector V. Specifically, the movement vector creation unit 323bh excludes, from the plurality of inside IPs, the inside IP whose unbiased variance is equal to or more than a predetermined value from the creation processing of the movement vector V.

  Returning to FIG. 2, the weighting unit 324b will be described. The weighting unit 324b weights each of the current instantaneous data 100 and the current particle data in the assignment relationship with respect to the current particle data.

  For example, the weighting unit 324b increases the weight of the particle data close to the current instant data 100 among the current particle data, and decreases the weight of the particle data far from the current instant data 100. Here, “close” and “far” mean that the Mahalanobis distance is “close” and “far”.

  The weighting unit 324b performs weighting based on the movement vector V estimated by the estimation unit 323b. Here, such a point will be described with reference to FIG. FIG. 10 is a diagram illustrating a weighting process based on the movement vector V.

  FIG. 10 shows a case where the starting points of the vectors PV1 to PV3 based on the three particle data and the starting point of the movement vector V are aligned with the origins of the VY-axis and VX-axis planes. The vectors PV1 to PV3 based on the particle data are vectors from the previous particle data to the current particle data (particle data predicted by the prediction unit 321b).

  For example, the weighting unit 324b increases the weight as the similarity between the vectors PV1 to PV3 and the movement vector V increases. For example, the weighting unit 324b increases the similarity as the ground speed components on the VY axis and the VX axis of the vectors PV1 to PV3 and the movement vector V are closer.

  That is, in the example shown in FIG. 10, the weighting unit 324b increases the weight of the particle data derived from the vectors PV1 and PV2 and decreases the weight of the particle data derived from the vector PV3. Thereby, the accuracy of resampling in the resampling unit 325b at the subsequent stage can be improved, and the accuracy of the calculated target data can be improved.

  Next, the resampling unit 325b rearranges (resamples) the particle data based on the weights of the current particle data. Specifically, the resampling unit 325b moves the particle data having a small weight closer to the instantaneous data 100.

  The target data generation unit 326b recalculates the probability density function based on the current particle data rearranged by the resampling unit 325b, and generates target data based on the center of gravity of the recalculated probability density function. . Although the target data generation unit 326b generates the target data based on the barycenter of the probability density function, the target data generation unit 326b may generate the target data based on, for example, an average of the probability density function.

  Further, the target data generation unit 326b treats the instantaneous data 100 to which the particle data has not been assigned as a new target, and outputs the data as it is as target data. That is, in the case of a new target, the target data generation unit 326b outputs the instantaneous data 100 = target data.

  Next, a processing procedure of processing executed by the target detection device 1 according to the embodiment will be described with reference to FIG. FIG. 11 is a flowchart illustrating a processing procedure of processing performed by the target detection device 1 according to the embodiment.

  As illustrated in FIG. 11, first, the generation unit 32a generates instantaneous data 100 including a relative speed of a direction of the target toward the own vehicle MC at each of a plurality of reflection points where the transmitted radio wave is reflected by the target. (Step S101).

  Subsequently, the prediction unit 321b of the filter processing unit 32b performs a prediction process of predicting the current particle data based on the previous particle data (Step S102). When there is no previous particle data, the filter processing unit 32b predicts the current particle data based on the relative speed of the previous instantaneous data 100.

  Subsequently, the allocating unit 322b allocates the current instantaneous data 100 to the current particle data (Step S103). Subsequently, the estimating unit 323b estimates the movement vector V of the target based on the plurality of current instant data 100 (step S104). The processing procedure of the estimation processing by the estimation unit 323b will be described later with reference to FIG.

  Subsequently, the assigning unit 322b determines the presence or absence of a new target based on the presence or absence of the instantaneous data 100 to which the current particle data has not been assigned (step S105). When the instantaneous data 100 is a new target (Step S105, Yes), the allocating unit 322b assigns predetermined particle data (for example, initial state particle data) to the instantaneous data 100 corresponding to the new target. Is set (step S106).

  Subsequently, when the instantaneous data 100 is not a new target (No at Step S105), the weighting unit 324b weights each of the current particle data based on the movement vector V estimated by the estimation unit 323b ( Step S107).

  Subsequently, the resampling unit 325b performs resampling of the current particle data based on the weight by the weighting unit 324b (Step S108). Subsequently, the target data generation unit 326b updates the probability density function of the resampled particle data of this time, generates target data based on the probability density function (step S109), and ends the process.

  Next, a processing procedure of an estimation process executed by the estimation unit 323b according to the embodiment will be described with reference to FIG. FIG. 12 is a flowchart illustrating a processing procedure of an estimation process performed by the estimation unit 323b according to the embodiment.

  As shown in FIG. 12, first, the instantaneous data selection unit 323ba performs a selection process of selecting the instantaneous data 100 used for the estimation process of the movement vector V from the plurality of instantaneous data 100 (step S201).

  Subsequently, the buffer update unit 323bb corrects the angle corresponding to the direction of the past instantaneous data 100 to the own vehicle MC in the history data 33a of the storage unit 33 based on the change amount of the rotation angle of the own vehicle MC. The update process of the history data 33a is performed (step S202).

  Subsequently, the ground speed calculation unit 323bc calculates the ground speed by removing the speed component of the own vehicle MC from the relative speed of the direction to the own vehicle MC in the instantaneous data 100 (step S203). Subsequently, the buffering unit 323bd performs a buffering process of storing the new instantaneous data 100 selected by the instantaneous data selection unit 323ba as the history data 33a (Step S204).

  Subsequently, the perpendicular creating unit 323be aligns the starting points in the ground vector GV corresponding to each of the plurality of instantaneous data 100 based on the ground speed in the direction to the host vehicle MC in the instantaneous data 100 calculated by the ground speed calculating unit 323bc. Then, a perpendicular line creating process for creating a perpendicular line VL with respect to the ground vector GV is performed (step S205).

  Subsequently, the intersection creating unit 323bf performs an intersection creating process of creating an intersection CP of the perpendicular VL created by the perpendicular creating unit 323be (Step S206). The movement vector creation unit 323bh determines whether or not there is one intersection CP created by the intersection creation unit 323bf (Step S207).

  When there is one intersection CP created by the intersection creation unit 323bf (step S207, Yes), the movement vector creation unit 323bh performs a movement vector creation process (step S208), and ends the process. In this case, the movement vector creation unit 323bh creates a vector from the starting point of the ground vector GV to the intersection CP as the movement vector V.

  On the other hand, in step S207, if the number of intersections CP is not one (step S207, No), that is, if there are three or more intersections CP, the inner center creation unit 323bg determines the inner center IP of the triangle formed by the three intersections CP. An inner core creation process to be created is performed (step S209), and the process proceeds to step S208. In such a case, in step S208, the movement vector creation unit 323bh performs a movement vector creation process based on the inner core IP created by the inner core creation unit 323bg.

  As described above, the target detection device 1 according to the embodiment includes the generation unit 32a, the ground speed calculation unit 323bc (an example of a calculation unit), and the estimation unit 323b. The generation unit 32a generates instantaneous data 100 including the relative speed of the direction of the target toward the own vehicle MC for each of the plurality of reflection points where the transmitted radio wave is reflected on the target. The ground speed calculating unit 323bc calculates the ground speed based on the relative speed of the direction to the own vehicle MC of the instantaneous data 100 generated by the generating unit 32a and the speed of the own vehicle MC. Based on the ground speed calculated by the ground speed calculation unit 323bca, the estimation unit 323b aligns the starting points in the ground vector GV of the direction to the own vehicle MC corresponding to each of the plurality of instantaneous data 100, and sets the vertical line VL to the ground vector GV. The movement direction of the target (one element of the movement vector V) is estimated from the intersection CP. Thereby, the response to the target can be improved.

  In the above-described embodiment, the target detection device 1 is provided in the vehicle. However, it is needless to say that the target detection device 1 may be provided in a moving object other than the vehicle, for example, a ship or an aircraft.

  Further, in the above-described embodiment, ESPRIT has been described as an example of the direction-of-arrival estimation method used by the target detection device 1, but the present invention is not limited to this. For example, DBF (Digital Beam Forming), PRISM (Propagator method based on an Improved Spatial-smoothing Matrix), MUSIC (Multiple Signal Classification), or the like may be used.

  The target detection device 1 according to the embodiment may verify the accuracy of the movement vector V created by the movement vector creation unit 323bh based on the particle data. This will be described with reference to FIG.

  FIG. 13 is a diagram illustrating the verification processing of the movement vector V. FIG. 13 shows a movement vector V and a vector PV based on the particle data.

  As illustrated in FIG. 13, the estimation unit 323b calculates the accuracy indicating the reliability of the movement vector V based on the similarity between the created movement vector V and the particle data vector PV. Specifically, the estimating unit 323b calculates the similarity between the ground speed components on the VY axis and the VX axis of the movement vector V and the vector PV as accuracy.

  For example, when the similarity between the particle data vector PV and the movement vector V is equal to or greater than a predetermined value, the estimation unit 323b sets the movement vector V as a final value of the final estimation result. On the other hand, when the similarity between the particle data vector PV and the movement vector V is less than a predetermined value, the estimation unit 323b prohibits the use of the movement vector V. As a result, it is possible to use the movement vector V with higher estimation accuracy.

  In FIG. 13, the similarity between the ground speed components on the VY axis and the VX axis of the movement vector V and the vector PV is calculated as the accuracy. May be calculated as

  In the above-described embodiment, the filter processing unit 32b shows the case where the particle filter is used. However, the time series filter to be used is not limited to the particle filter. For example, a Kalman filter, an αβ filter, or the like is used. It may be a time series filter.

  Further effects and modifications can be easily derived by those skilled in the art. Thus, the broader aspects of the present invention are not limited to the specific details and representative embodiments shown and described above. Accordingly, various modifications can be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.

Reference Signs List 1 target detection device 32 signal processing unit 32a generation unit 32b filter processing unit 33 storage unit 33a history data 100 instantaneous data 321b prediction unit 322b allocation unit 323b estimation unit 323ba instantaneous data selection unit 323bb buffer update unit 323bc buffer for ground speed calculation 323bbd Ring part 323be Perpendicular line creation part 323bf Intersection creation part 323bg Inner center creation part 323bh Movement vector creation part 324b Weighting part 325b Resampling part 326b Target data generation part LC Other vehicle MC Own vehicle

Claims (9)

For each of the plurality of reflection points of the transmitted radio wave reflected by the target, a generation unit that generates instantaneous data including a relative speed of the direction of the target toward the own vehicle,
A calculating unit that calculates a ground speed based on the relative speed of the instantaneous data generated by the generating unit and the speed of the host vehicle;
Based on the ground speed calculated by the calculation unit, the starting points in the ground vector of the direction to the host vehicle corresponding to each of the plurality of instantaneous data are aligned, and the moving direction of the target from the intersection of the perpendicular to the ground vector. A target detecting device, comprising: an estimating unit for estimating. The estimation unit includes
The target detecting device according to claim 1, wherein when the number of the intersections is one, the direction from the starting point to the intersection in the ground vector is estimated to be the moving direction. The estimation unit includes
3. The target detection device according to claim 1, wherein when the number of the intersections is three or more, the movement direction is estimated based on a center of a triangle formed by the intersections. 4. The estimation unit includes
The method according to any one of claims 1 to 3, wherein, among the plurality of instantaneous data, the instantaneous data in which the direction of the ground vector toward the own vehicle is similar is excluded from the movement direction estimation processing. The target detection device described in the above. The instantaneous data includes
The distance from the own vehicle to the position of the instantaneous data is included,
The estimation unit includes
The instantaneous data having a similar relative speed and a similar distance among the plurality of instantaneous data generated by the generation unit are excluded from the estimation process of the moving direction. 4. The target detection device according to any one of 4. A storage unit that stores the instantaneous data generated by the generation unit as history data,
The estimation unit includes
When the number of the latest instantaneous data generated by the generation unit is less than a predetermined number, the past instantaneous data stored as the history data is used for the estimating process of the moving direction. Item 7. The target detection device according to any one of Items 1 to 5. The estimation unit includes
The target detection device according to claim 6, wherein the direction of the own vehicle in the past instantaneous data in the history data is corrected based on a change in the direction of the own vehicle. A filter processing unit that generates target data corresponding to the instantaneous data by applying a particle filter that assigns a plurality of particle data to the instantaneous data generated by the generation unit,
The estimation unit includes
When the degree of similarity between the direction of the vector based on the plurality of particle data and the movement direction is equal to or greater than a predetermined value, the movement direction is used as the final value of the estimation result. The target detection device according to any one of the above. For each of the plurality of reflection points of the transmitted radio wave reflected by the target, a generation step of generating instantaneous data including a relative speed of the direction of the target toward the own vehicle,
A calculating step of calculating a ground speed based on the relative speed of the instantaneous data generated by the generating step and the speed of the host vehicle;
Based on the ground speed calculated in the calculating step, the starting points in the ground vector of the direction to the host vehicle corresponding to each of the plurality of instantaneous data are aligned, and the moving direction of the target from the intersection of the perpendicular to the ground vector. An estimation step of estimating the target.

Images