JP2018080938A - Radar device and target detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a radar device and a target detection method with which it is possible to continuously track a target that could be a pedestrian.SOLUTION: A radar device according to an embodiment comprises a detection unit, an extrapolation unit, a determination unit and a candidate setting unit. The detection unit repeatedly performs a detection process for detecting a target approaching the own vehicle. The extrapolation unit performs an extrapolation process for predicting a virtual position of the target that can no longer be detected by the latest detection process on the basis of the result of the previous detection process. The determination unit performs a determination process for determining in sequence of time whether or not the same target that became the object of the detection process or the extrapolation process is a pedestrian who could collide with the own vehicle. When the target that became the object of the extrapolation process was determined to be a pedestrian in the previous determination process, the candidate setting unit sets the target to a candidate to be continuously tracked as a pedestrian without excluding it from candidates to be tracked as a pedestrian.SELECTED DRAWING: Figure 1

Description

  Embodiments disclosed herein relate to a radar apparatus and a target detection method.

  Conventionally, there is a radar device that receives a reflected wave of a transmission wave transmitted to a target existing ahead of the host vehicle and detects the target based on a reflected wave whose signal intensity exceeds a threshold (for example, Patent Document 1). Such on-vehicle radar devices have been used to detect preceding vehicles and oncoming vehicles, but in recent years, their use has expanded to detect pedestrians.

  When the radar device detects a pedestrian, the signal intensity of the reflected wave from the pedestrian is lower than the signal intensity of the reflected wave from the other vehicle. The threshold used for detection needs to be set low.

JP 2012-013484 A

  However, the radar apparatus may erroneously detect a non-existent target due to noise emitted from other electronic devices mounted on the host vehicle. In order to eliminate such noise, there is a risk of misdetecting the noise as a target. For some areas, the threshold could not be set low. For this reason, it has been difficult for the conventional radar apparatus to continuously track a target that may be a pedestrian when the distance between the pedestrian and the host vehicle is a certain distance.

  One aspect of the embodiments has been made in view of the above, and an object thereof is to provide a radar device and a target detection method capable of continuously tracking a target that may be a pedestrian. To do.

  A radar apparatus according to an aspect of an embodiment includes a detection unit, an extrapolation unit, a determination unit, and a candidate setting unit. The detection unit repeatedly performs a detection process for detecting a target approaching the host vehicle. The extrapolation unit performs an extrapolation process for predicting the virtual position of the target that is no longer detected in the current detection process based on the result of the previous detection process. The determination unit performs a determination process for determining, in a time series, whether or not the same target that is the target of the detection process or the extrapolation process is a pedestrian that may collide with the host vehicle. . The candidate setting unit excludes the target from the tracking target candidates as the pedestrian when the target that has been subjected to the extrapolation processing has been determined as the pedestrian in the previous determination process. Without being set as a tracking target candidate as the pedestrian.

  According to the radar apparatus and the target detection method according to an aspect of the embodiment, a target that may be a pedestrian can be continuously tracked.

FIG. 1 is an explanatory diagram illustrating a target detection method according to the embodiment. FIG. 2 is a block diagram illustrating the radar apparatus according to the embodiment. FIG. 3A is an explanatory diagram of processing from pre-processing in the signal processing device according to the embodiment to peak extraction processing in the signal processing device. FIG. 3B is an explanatory diagram of processing from pre-processing in the signal processing device according to the embodiment to peak extraction processing in the signal processing device. FIG. 3C is an explanatory diagram of processing from pre-processing of the signal processing device according to the embodiment to peak extraction processing in the signal processing device. FIG. 4A is a process explanatory diagram of the azimuth calculation process according to the embodiment. FIG. 4B is a process explanatory diagram (part 1) of the pairing process according to the embodiment. FIG. 4C is a process explanatory diagram (part 2) of the pairing process according to the embodiment. FIG. 5 is an explanatory diagram illustrating threshold values used for the peak extraction process according to the embodiment. FIG. 6 is a block diagram illustrating a target classification unit according to the embodiment. FIG. 7A is an explanatory diagram illustrating an example of movement conditions according to the embodiment. FIG. 7B is an explanatory diagram illustrating an example of the movement condition according to the embodiment. FIG. 8 is an operation explanatory diagram of the target classification unit according to the embodiment. FIG. 9 is a block diagram illustrating the peak extraction unit according to the embodiment. FIG. 10 is an explanatory diagram illustrating an example of the noise condition according to the embodiment. FIG. 11 is a block diagram illustrating a count processing unit according to the embodiment. FIG. 12 is an explanatory diagram illustrating an example of a counter operation condition according to the embodiment. FIG. 13 is a flowchart illustrating main processing executed by the data processing unit of the radar apparatus according to the embodiment. FIG. 14 is a flowchart illustrating peak extraction processing during main processing according to the embodiment. FIG. 15 is a flowchart showing the target classification process in the main process according to the embodiment. FIG. 16 is a flowchart showing the counting process during the main process according to the embodiment. FIG. 17 is an explanatory diagram illustrating an example of a processing result obtained by the radar apparatus according to the embodiment.

  Hereinafter, embodiments of a radar apparatus and a target detection method disclosed in the present application will be described in detail with reference to the accompanying drawings. In addition, this invention is not limited by embodiment shown below.

  First, a target detection method performed by the radar apparatus according to the embodiment will be described with reference to FIG. FIG. 1 is an explanatory diagram illustrating a target detection method according to the embodiment. Here, the case where a radar apparatus is provided in the center of the front grille of the own vehicle C is demonstrated.

  Note that the position where the radar device is provided is not limited to this, and may be an arbitrary position such as a rear portion, a side surface, or a rearview mirror of the host vehicle C. Hereinafter, the traveling direction of the host vehicle C may be referred to as a vertical direction, and the direction intersecting the traveling direction of the host vehicle C may be referred to as a lateral direction.

  As shown in FIG. 1, the radar device is a device that detects a target in a detection region A that spreads a predetermined angle from the front of the host vehicle C to the left and right, for example, using the radar device as a key of a fan. Such a radar apparatus receives a reflected wave reflected from a target transmitted by the transmission wave transmitted toward the detection area A, and repeats a detection process for detecting the target based on a reflected wave whose signal intensity exceeds a threshold value. The target is detected continuously in time series.

  In addition, the radar apparatus may not detect the target by the detection process when the target is outside the detection area A, or when the signal intensity of the reflected wave is equal to or lower than the threshold even if the target is within the detection area A. May be possible. In such a case, the radar apparatus performs extrapolation processing for predicting the virtual position of the target that is no longer detected in the current detection processing based on the result of the previous detection processing.

  The radar apparatus 1 also includes a counter (hereinafter referred to as “survival counter”) that counts for each target a count value indicating the high possibility that the target that is the target of detection processing or extrapolation processing actually exists. To be described). The survival counter adds a predetermined value to the count value every time the target is detected by the detection process, and subtracts the predetermined value from the count value every time the virtual position of the target is predicted by the extrapolation process. It is an example of a counter part.

  For example, as shown in FIG. 1, when the radar apparatus detects a pedestrian P in the detection area A at time t1, the radar apparatus adds a predetermined value to the count value of the pedestrian P. Thereafter, when the pedestrian P leaves the detection area A at time t2, the radar apparatus performs extrapolation processing, and therefore subtracts a predetermined value from the count value of the pedestrian P.

  The radar apparatus sets a target whose count value of the survival counter is larger than a predetermined threshold (for example, 0) as a tracking target candidate as a pedestrian who may collide with the own vehicle C. Then, the radar apparatus excludes the target whose count value of the survival counter is a predetermined threshold (for example, 0) or less from the tracking target candidates as pedestrians that may collide with the host vehicle C.

  In addition, when the pedestrian P exists in the detection area A and can be detected by the detection process at time t2, the radar apparatus further adds a predetermined value to the count value. At this time, when the pedestrian P is stationary between the time t1 and the time t2, when the host vehicle C moves forward, the vertical distance between the pedestrian P and the host vehicle C decreases, but the pedestrian P The lateral distance between the vehicle and the host vehicle C does not change. For this reason, the radar apparatus performs a determination process for determining the pedestrian P as a target having a low possibility of colliding with the own vehicle C.

  Thereafter, when the pedestrian P moves in the lateral direction across the traveling direction of the host vehicle C between the times t2 and t3, the longitudinal and lateral distances between the pedestrian P and the host vehicle C are shortened. In this way, the radar apparatus determines that the pedestrian P is likely to collide with the own vehicle C when the lateral movement of the pedestrian P crossing the traveling direction of the own vehicle C is detected. Process. The radar device further adds a predetermined value to the count value when a pedestrian is present in the detection area A at time t3 and can be detected by the detection process.

  Thereafter, when the pedestrian P continues to move in the lateral direction across the traveling direction of the host vehicle C and the host vehicle C continues to travel, the pedestrian P may leave the detection area A at time t4. In this case, since a general radar apparatus performs extrapolation processing, a predetermined value is subtracted from the count value. Thereby, the general radar apparatus may exclude the pedestrian P from the tracking target candidates as pedestrians that may collide with the host vehicle C, and may not be able to continuously track.

  Therefore, the radar apparatus according to the embodiment has determined that there is a possibility that the pedestrian P subject to extrapolation processing at time t4 may collide with the own vehicle C in the previous determination processing. For the pedestrian P, instead of subtracting the count value, the current count value is held in the survival counter.

  In this way, the radar apparatus does not exclude the pedestrian P from the tracking target candidate as a pedestrian who may collide with the own vehicle C, and the tracking target as a pedestrian who may continue to collide with the own vehicle C. Set as a candidate. Thereby, the radar apparatus according to the embodiment can continuously track a target that may be a pedestrian.

  Next, the configuration of the radar apparatus 1 according to the embodiment will be described with reference to FIG. FIG. 2 is a block diagram illustrating the radar apparatus 1 according to the embodiment. In FIG. 2, only components necessary for explaining 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 functionally conceptual 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 one shown in the figure, and all or a part thereof is functionally or physically distributed in arbitrary units according to various loads or usage conditions.・ It can be integrated and configured.

  As shown in FIG. 2, the radar apparatus 1 includes a transmission unit 2 and a transmission antenna 4 as components that constitute a transmission system. The transmission unit 2 includes a signal generation unit 21 and an oscillator 22.

  In addition, the radar apparatus 1 includes reception antennas 5-1 to 5-n and reception units 6-1 to 6-n as components constituting the reception system. Each of the reception units 6-1 to 6-n includes a mixer 61 and an A / D conversion unit 62. The radar apparatus 1 includes a signal processing device 7 as a component constituting the signal processing system.

  In the following, for the sake of simplification of explanation, when “reception antenna 5” is simply described, the reception antennas 5-1 to 5-n are collectively referred to. The same applies to the “receiving unit 6”.

  The transmission unit 2 performs processing for generating a transmission signal. The signal generation unit 21 generates a modulation signal for transmitting a millimeter wave that is frequency-modulated with a triangular wave, under the control of a transmission / reception control unit 71 provided in the signal processing device 7 described later. The oscillator 22 generates a transmission signal based on the modulation signal generated by the signal generation unit 21.

  The transmission antenna 4 transmits the transmission signal generated by the oscillator 22 as a transmission wave in front of the host vehicle C. As shown in FIG. 2, the transmission signal generated by the oscillator 22 is also distributed to the mixer 61 described later.

  The reception antenna 5 receives the reflected wave coming from the target as a reception signal by reflecting the transmission wave transmitted from the transmission antenna 4 on the target. Each of the reception units 6 performs pre-processing until each received signal is passed to the signal processing device 7.

  Specifically, each of the mixers 61 mixes the transmission signal distributed as described above and the reception signal received at each of the reception antennas 5 to generate a beat signal. A corresponding amplifier may be provided between the receiving antenna 5 and the mixer 61.

  The A / D converter 62 digitally converts the beat signal generated in the mixer 61 and outputs it to the signal processing device 7. The signal processing device 7 includes a transmission / reception control unit 71, an FFT (Fast Fourier Transform) unit 72, a data processing unit 73, and a storage unit 74.

  The data processing unit 73 includes a peak extraction unit 73a, an azimuth calculation unit 73b, a pairing unit 73c, a continuity determination unit 73d, a filter processing unit 73e, a count processing unit 73f, a target classification unit 73g, an unnecessary target determination unit 73h, A combination processing unit 73i and an output target selection unit 73j are provided.

  The storage unit 74 is a storage device such as a hard disk drive, a nonvolatile memory, or a register, and stores a noise condition 74a, a counter operation condition 74b, a count value 74c, a movement condition 74d, and target history information 74e.

  The noise condition 74a is information used when the peak extraction unit 73a performs a noise determination process described later. The counter operation condition 74b is information referred to when the count processing unit 73f performs a count process described later. The count value 74c is a value counted for each target by the aforementioned survival counter. The movement condition 74d is information that is referred to when the target classification unit 73g performs movement determination described later.

  The target history information 74e is information indicating a history regarding the position and the like of the target classified by the target classification unit 73g by a target classification process described later. In other words, it is information indicating a trajectory or the like in which the same target has moved continuously in time. The information other than the position of the target included in the history is, for example, information indicating a relative speed between the radar apparatus 1 and the target, a signal level at a peak frequency of a beat signal described later, and the like.

  The transmission / reception control unit 71 controls the transmission unit 2 including the signal generation unit 21 described above. Further, although not shown, the control of each receiving unit 6 is also performed. The FFT unit 72 performs fast Fourier transform on the beat signal input from each A / D conversion unit 62 and outputs the result to the peak extraction unit 73 a of the data processing unit 73.

  The peak extraction unit 73a extracts a peak frequency that becomes a peak in the fast Fourier transform result by the FFT unit 72 and outputs the peak frequency to the azimuth calculation unit 73b. The peak extraction unit 73a extracts a peak frequency for each of an “UP section” and a “DN section” of a beat signal to be described later.

  Further, the peak extraction unit 73a performs noise determination processing determination as to whether or not the extracted peak frequency is EMC (Electro Magnetic Compatibility) noise, and turns on the EMC noise determination flag when determining that the extracted peak frequency is EMC noise. Then, the peak extraction unit 73a outputs information indicating the state of the EMC noise determination flag to the count processing unit 73f. The peak extraction unit 73a will be described in detail later with reference to FIGS.

  The azimuth calculation unit 73b calculates the arrival angle of the reflected wave corresponding to each of the peak frequencies extracted by the peak extraction unit 73a and its signal intensity (reception level). At this time, the arrival angle includes the case where the phase is turned back and is an angle at which the target is estimated to exist. In addition, the azimuth calculation unit 73b outputs the calculated estimated angle and reception level to the pairing unit 73c.

  The pairing unit 73c determines the correct combination of the peak frequencies of the “UP section” and the “DN section” based on the calculation result of the azimuth calculation unit 73b, and calculates the distance and relative speed of each target from the combination result. . The pairing unit 73c outputs information including the estimated angle, distance, and relative speed of each target to the continuity determination unit 73d.

  Here, the flow of a series of processes so far in the signal processing device 7 will be described with reference to FIGS. 3A, 3B, 3C, 4A, 4B, and 4C. 3A, 3B, and 3C are explanatory diagrams of processing from the previous stage processing of the signal processing device 7 to the peak extraction processing in the signal processing device 7 according to the embodiment.

  FIG. 4A is a process explanatory diagram of the azimuth calculation process according to the embodiment. FIG. 4B is a process explanatory diagram (part 1) of the pairing process according to the embodiment. FIG. 4C is a process explanatory diagram (No. 2) of the pairing process according to the embodiment.

  As shown in FIG. 3A, the transmission signal fs (t) is transmitted from the transmission antenna 4 as a transmission wave, then reflected by the target and arrives as a reflection wave, and received at the reception antenna 5 as a reception signal fr (t). Received.

  At this time, as shown in FIG. 3A, the reception signal fr (t) is delayed by a time difference τ with respect to the transmission signal fs (t) according to the distance between the host vehicle C and the target. Due to this time difference τ and the Doppler effect based on the relative speed of the host vehicle C and the target, the frequency increases in the output signal obtained by mixing the reception signal fr (t) and the transmission signal fs (t). A beat signal is obtained in which the frequency “up” of the “UP section” and the frequency fdn of the “DN section” where the frequency decreases are repeated.

  FIG. 3B schematically shows the result of fast Fourier transform of the beat signal in the FFT unit 72 for the “UP section” side. FIG. 3C schematically shows the result of fast Fourier transform of the beat signal in the FFT unit 72 for the “DN section” side.

  As shown in FIGS. 3B and 3C, after the fast Fourier transform, waveforms in the respective frequency domains on the “UP section” side and the “DN section” side are obtained. The peak extraction unit 73a extracts a peak frequency at which the signal intensity has a maximum value (peak) in the waveform.

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

  Further, as shown in FIG. 3C, on the “DN section” side, the peaks Pd1, Pd2, and Pd3 are determined as peaks by the peak extraction threshold, and the peak frequencies fd1, fd2, and fd3 are extracted.

  Here, the frequency component of each peak frequency extracted by the peak extraction unit 73a may be a mixture of reflected waves from a plurality of targets. Therefore, the azimuth calculation unit 73b performs azimuth calculation for each peak frequency and analyzes the presence of a target corresponding to each peak frequency.

  Note that the azimuth calculation in the azimuth calculation unit 73b is performed using a predetermined arrival direction estimation method such as ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques), but is not limited thereto.

  FIG. 4A schematically shows the result of the azimuth calculation performed by the azimuth calculation unit 73b. The azimuth calculation unit 73b calculates an estimated angle of each target corresponding to each of the peaks Pu1 to Pu3 from the peaks Pu1 to Pu3 of the azimuth calculation result. Moreover, the magnitude | size of each peak Pu1-Pu3 becomes a reception level. The azimuth calculation unit 73b performs the azimuth calculation processing on each of the “UP section” side and the “DN section” side.

  As shown in FIG. 4B, the pairing unit 73c performs pairing that combines the peaks with the estimated angles and the reception levels close to each other in the azimuth calculation result of the azimuth calculation unit 73b. Further, from the combination result, the pairing unit 73c calculates the distance and relative velocity of each target corresponding to each peak combination. The distance is calculated based on the relationship of “distance ∝ (fup + fdn)”. The relative speed is calculated based on the relationship of “speed ∝ (fup−fdn)”.

  Thus, as shown in FIG. 4C, the pairing unit 73c obtains a pairing processing result indicating the estimated angle, distance, and relative speed of each target TG with respect to the host vehicle C, thereby detecting each target TG. It functions as a detection unit that performs processing. And the pairing part 73c outputs the pairing process result used as a detection process result to the continuity determination part 73d.

  Returning to the description of FIG. 2, the continuity determination unit 73d will be described. The continuity determination unit 73d determines whether or not the instantaneous value of the target being determined by the current scan has continuity with the target detected by the previous scan. Specifically, the current predicted position is calculated from the position of the target up to the previous scan, and if there is an instantaneous value close to the predicted position in the current scan, it is determined that the instantaneous value is continuous. Then, the continuity determination unit 73d outputs information on the target after the continuity determination to the filter processing unit 73e.

  The filter processing unit 73e is a processing unit that corrects variation in instantaneous values for each target by a filtering process that averages the instantaneous values for a plurality of times processed in time series. The filter processing unit 73e outputs information on the target after the filtering process to the target classification unit 73g.

  Further, when there is no instantaneous value close to the predicted position in the current detection process as a result of the continuity determination, the filter processing unit 73e detects the current detection based on the instantaneous value of the same target detected in the previous detection process. An extrapolation process for predicting the virtual position of the target that is no longer detected by the process is performed.

  When the extrapolation process is performed, the filter processing unit 73e turns on the extrapolation flag of the target whose virtual position is predicted by the extrapolation process. Moreover, the filter process part 73e turns off an extrapolation flag, when an extrapolation process is not performed.

  Then, the filter processing unit 73e outputs information indicating the state of the extrapolation flag to the count processing unit 73f. In this way, the filter processing unit 73e functions as an extrapolation unit that performs extrapolation processing for predicting the virtual position of a target that is no longer detected in the current detection processing based on the target detected in the previous detection processing. To do.

  When the information indicating that the extrapolation flag is OFF is input from the filter processing unit 73e, the count processing unit 73f adds a predetermined value to the count value 74c of the aforementioned survival counter provided therein, and stores the storage unit 74. Remember me. In addition, when the information indicating that the extrapolation flag is ON is input from the filter processing unit 73e, the count processing unit 73f subtracts a predetermined value from the count value 74c of the survival counter and stores it in the storage unit 74. .

  The target classification unit 73g classifies each target into a moving object (for example, a preceding vehicle, an oncoming vehicle, a pedestrian P) and a stationary object based on the filter processing result of the filter processing unit 73e. The target classification unit 73g outputs the classified classification result to the unnecessary target determination unit 73h.

  The unnecessary target determination unit 73h determines whether or not the target is unnecessary in terms of system control. Unnecessary targets are structures, wall reflections, and the like. In addition, although the target made unnecessary is not made into the output object to an external apparatus fundamentally, you may hold | maintain internally.

  The unnecessary target determination unit 73h refers to the count value 74c for each target stored in the storage unit 74, and determines that the target for which the count value 74c is “0” is also unnecessary. Then, the unnecessary target determination unit 73h outputs information related to the target that is not determined to be unnecessary to the combination processing unit 73i.

  The combination processing unit 73i performs grouping of a plurality of targets that are detected as being present and that are presumed to be reflection points from the same object, into one target.

  The output target selection unit 73j selects a target that needs to be output to an external device for system control. In addition, the output target selection unit 73j outputs target information regarding the selected target (including the actual angle, distance, relative speed, etc.) to the external device.

  Here, the external device is, for example, the vehicle control device 10. The vehicle control device 10 is an ECU (Electronic Control Unit) that controls each device of the host vehicle C. The vehicle control device 10 is electrically connected to, for example, a vehicle speed sensor 11, a rudder angle sensor 12, a throttle 13, and a brake 14.

  The vehicle control device 10 performs vehicle control such as ACC (Adaptive Cruise Control) and PCS (Pre-Crash Safety System) based on the target information acquired from the radar device 1.

  For example, when ACC is performed, the vehicle control device 10 uses the target information acquired from the radar device 1 so that the host vehicle C follows the preceding vehicle while keeping the inter-vehicle distance from the preceding vehicle constant. The throttle 13 and the brake 14 are controlled. In addition, the vehicle control device 10 obtains the traveling state of the host vehicle C that changes from time to time, that is, the vehicle speed, the steering angle, and the like from the vehicle speed sensor 11, the steering angle sensor 12, etc., and feeds back to the radar device 1.

  Further, for example, when performing the PCS, the vehicle control device 10 uses the target information acquired from the radar device 1 and there is a preceding vehicle, a stationary object, or the like that has a risk of collision in the traveling direction of the host vehicle C. If this is detected, the brake 14 is controlled to decelerate the host vehicle C. In addition, for example, an alarm device (not shown) is used to warn the passenger of the host vehicle C, or the passenger is fixed to the seat by retracting a seat belt in the passenger compartment.

  Since the radar apparatus 1 may erroneously detect a target that does not actually exist in the vicinity of the own vehicle C due to noise emitted from other electronic devices such as a navigation device mounted on the own vehicle C, the vicinity of the own vehicle C For these areas, the peak extraction threshold is set high.

  Here, the peak extraction threshold according to the embodiment will be described with reference to FIG. FIG. 5 is an explanatory diagram illustrating a peak extraction threshold according to the embodiment. As illustrated in FIG. 5, the radar apparatus 1 sets a high peak extraction threshold value in a frequency region of 5 [BIN] or less, for example, according to a frequency region in which EMC noise is generated. Here, the interval between the BINs is a frequency interval corresponding to the frequency resolution of the FFT unit 72.

  Thereby, the radar apparatus 1 can prevent erroneous detection of a target that does not actually exist in the vicinity of the host vehicle C due to EMC noise emitted from other electronic devices mounted on the host vehicle C. .

  However, when the pedestrian P exists in an area where the peak extraction threshold is set relatively high, that is, in an area close to the own vehicle C, the radar apparatus 1 has a low signal intensity compared to other vehicles. In some cases, the peak of the reflected wave from P cannot be extracted.

  The noise frequency region described so far is merely an example. Therefore, the radar apparatus 1 can set a high peak extraction threshold in a specific area other than the area close to the host vehicle C according to the frequency of noise output from another electronic device. Thereby, the radar apparatus 1 can prevent erroneous detection of a target that does not actually exist at a distance corresponding to a specific area.

  Therefore, the radar apparatus 1 can collide with the host vehicle C by distinguishing noises emitted from other electronic devices mounted on the host vehicle C and pedestrians P located near the host vehicle C. The structure which tracks the pedestrian P with the nature continuously is provided.

  Specifically, the target classification unit 73g may cause the target detected by the pairing unit 73c and the target whose virtual position is predicted by extrapolation processing by the filter processing unit 73e to collide with the host vehicle C. The structure which determines whether it is a pedestrian P with the nature is provided.

  Moreover, the peak extraction part 73a is equipped with the structure which determines whether the peak detected this time is the noise emitted from the other electronic device mounted in the own vehicle C. FIG. In addition, the count processing unit 73f has a configuration for holding the count value 74c of the survival counter when the target that has been subjected to the extrapolation processing this time is determined to be the pedestrian P in the previous determination processing.

  Hereinafter, the noise condition 74a, the counter operation condition 74b, the movement condition 74d, and the target history information 74e will be described together with the configurations and operations of the target classification unit 73g, the peak extraction unit 73a, and the count processing unit 73f.

  FIG. 6 is a block diagram illustrating the target classification unit 73g according to the embodiment. 7A and 7B are explanatory diagrams illustrating an example of the movement condition 74d according to the embodiment. FIG. 8 is an operation explanatory diagram of the target classification unit 73g according to the embodiment.

  As illustrated in FIG. 6, the target classification unit 73g includes a classification unit 75a and a determination unit 75b. The classification unit 75a is a processing unit that classifies each target into a moving object and a stationary object based on information regarding the target input from the filter processing unit 73e.

  The moving object is, for example, a preceding vehicle, an oncoming vehicle, a walking pedestrian P, or the like. The stationary object is, for example, a roadside object such as a guardrail or a sign, a structure such as a building, or a pedestrian P that is stopped. The classification unit 75a outputs the target classification result to the unnecessary target determination unit 73h, and outputs information including the position of the target classified as the pedestrian P to the determination unit 75b.

  The determination unit 75b is a processing unit that determines whether or not the pedestrian P is likely to collide with the host vehicle C using the characteristics of the pedestrian P. Specifically, the pedestrian P who may collide with the own vehicle C has a feature of moving in the lateral direction intersecting the traveling direction of the own vehicle C.

  Therefore, when information related to the target is input from the classification unit 75a, the determination unit 75b first displays information indicating the position of the target included in the information (hereinafter referred to as “target position”). It adds to the target history information 74e.

  Subsequently, the determination unit 75b determines whether the target position added before this time already exists in the target history information 74e for the target to which the target position is added. And when the target position added before this time already exists, the determination part 75b is based on the target position added before this time, the target position added this time, and the movement condition 74d. Thus, it is determined whether or not the target is a pedestrian P that may collide with the host vehicle C.

  The movement condition 74d includes, for example, a movement condition for temporary determination of the lateral position movement amount shown in FIG. 7A and a movement condition for determination of the lateral position movement amount shown in FIG. 7B. In the pedestrian area determination flag shown in FIG. 7A, when the target position added this time to the target history information 74e is in the pedestrian area, the determination unit 75b is turned ON, and the target position added this time is walking. This flag is turned off by the determination unit 75b when not in the person area.

  The pedestrian area here is, for example, an area within 2.5 [m] on the left and right of the host vehicle C within 40 [m] in front of the host vehicle C. In addition, the area of this pedestrian area is an example, and is not limited to this.

  A curve R shown in FIG. 7A is the radius of an arc drawn by the road on which the host vehicle C is traveling. In the present embodiment, the curve R when the road is a right curve is positive, and the curve R when the road is a left curve is negative.

  Further, the movement start lateral position shown in FIG. 7A is a coordinate in the lateral direction with respect to the radar apparatus 1 of the target position when the target starts moving in a direction intersecting the traveling direction of the host vehicle C. Further, the current horizontal position shown in FIGS. 7A and 7B is a coordinate in the horizontal direction with respect to the radar apparatus 1 of the target position added this time to the target history information 74e.

  Further, the movement start lateral position shown in FIG. 7B is a coordinate in the lateral direction of the target position with respect to the radar apparatus 1 when the movement condition for the lateral position movement amount provisional determination is satisfied. In the present embodiment, the horizontal coordinate on the right side of the radar apparatus 1 is positive, and the horizontal coordinate on the left side of the radar apparatus 1 is negative.

  In addition, the movement start vertical position shown in FIG. 7B is the vertical direction of the target with respect to the radar device 1 (a direction parallel to the traveling direction of the host vehicle C) when the movement condition 74d for the horizontal position movement amount provisional determination is satisfied. Coordinates. The current vertical position is a coordinate in the vertical direction (a direction parallel to the traveling direction of the host vehicle C) with respect to the radar apparatus 1 of the target position added to the current target history information 74e.

  When the movement condition 74d for the horizontal position movement amount temporary determination shown in FIG. 7A is satisfied, the determination unit 75b turns on the horizontal position movement amount temporary determination flag shown in FIG. 7B and sets the horizontal position movement amount temporary determination shown in FIG. 7A. When the determination moving condition 74d is not satisfied, the horizontal position movement amount temporary determination flag shown in FIG. 7B is turned OFF.

  The condition shown in FIG. 7A is a condition for the radar apparatus 1 to detect a target that moves in a direction crossing the traveling direction of the vehicle C. Specifically, when the pedestrian area determination flag is ON and the absolute value of the curve R is 1000 [m] or more, the determination unit 75b determines that one of the following two conditions is satisfied. Then, the lateral position movement amount temporary determination flag is turned ON.

  The first condition is that the movement start lateral position is 0 [m] or more, and the value obtained by subtracting the current lateral position from the movement start lateral position is 0.5 [m] or more, which is an example of the first threshold. It is. That is, the first condition is a condition that a target existing on the right side of the radar apparatus 1 has approached the radar apparatus 1 by 0.5 [m] or more in the horizontal direction.

  The second condition is that the movement start lateral position is smaller than 0 [m], and the value obtained by subtracting the movement start lateral position from the current lateral position is 0.5 [m] or more, which is an example of the first threshold. It is a condition. That is, the second condition is a condition that the target existing on the left side of the radar apparatus 1 has approached the radar apparatus 1 by 0.5 [m] or more in the horizontal direction.

  Thus, the determination unit 75b provisionally determines the target as a pedestrian P when the target moves in a direction that intersects the traveling direction of the host vehicle C. Thereby, the determination part 75b can tentatively determine whether or not the target is a pedestrian P who may collide with the host vehicle C.

  In addition, the determination unit 75b determines that the target is a pedestrian (hereinafter, simply referred to as “pedestrian”) that may collide with the host vehicle C when a predetermined movement condition 74d is satisfied. The moving condition 74d includes that the distance that the target has moved in the direction intersecting the traveling direction of the host vehicle C is equal to or greater than the first threshold value. Thereby, the determination part 75b can tentatively determine whether or not the target is a pedestrian P that may collide with the host vehicle C.

  Furthermore, the determination unit 75b determines that the target is a pedestrian P that may collide with the host vehicle C when the movement condition 74d for determining the horizontal position movement amount shown in FIG. Turn on the movement amount determination flag. Further, the determination unit 75b determines that the target is not a pedestrian P that may collide with the host vehicle C when the movement condition 74d for determining the lateral position movement amount shown in FIG. The position movement amount determination flag is turned OFF.

  The condition shown in FIG. 7B is a condition for detecting a pedestrian target that may collide with the host vehicle C. Specifically, when the lateral position movement amount temporary determination flag is ON and the absolute value of the curve R is 1000 [m] or more, the determination unit 75b satisfies one of the following two conditions. In this case, the lateral position movement amount provisional determination flag is turned ON.

  The first condition is that the movement start lateral position is 0 [m] or more, and a value obtained by subtracting the current lateral position from the movement start lateral position is 0.7 [m] or more as an example of the second threshold. The condition is that the distance from the movement start vertical position to the current vertical position is Y [m] or less, which is an example of the third threshold value.

  That is, the first condition is that the target on the right side of the radar apparatus 1 is 0.7 [m] or more in the lateral direction from the time when the movement condition 74d for the lateral position movement amount temporary determination is satisfied. The distance that the vehicle C traveled during that time is Y [m] or less.

  The second condition is that the movement start lateral position is smaller than 0 [m], and the value obtained by subtracting the movement start lateral position from the current lateral position is 0.7 [m] or more, which is an example of the second threshold. There is a condition that the distance from the movement start vertical position to the current vertical position is equal to or less than Y [m], which is an example of the third threshold.

  In other words, the second condition is that the target existing on the left side of the radar apparatus 1 is 0.7 [m] or more in the lateral direction from the time when the movement condition 74d for temporary determination of the lateral position movement amount is satisfied. The distance that the vehicle C traveled during that time is Y [m] or less. Here, Y [m] is a distance that varies depending on the speed of the host vehicle C, and decreases as the speed of the host vehicle C decreases. For example, when the speed of the vehicle C is 60 [km / h], Y is 20 [m].

  The determination unit 75b then walks where the target may collide with the host vehicle C based on the moving condition 74d, the target position added before this time, and the target position added this time. It is determined whether or not the person P.

  For example, as illustrated in FIG. 8, the determination unit 75b determines that the target TG exists in the pedestrian area on the left side of the host vehicle C at time t11, and then the target TG is between time t12 and time t13. After moving 0.5 [m] to the right, the lateral position movement amount temporary determination flag is turned ON.

  Further, the determination unit 75b then moves the target TG 0.7 [m] to the right during the period from time t13 to time t14, during which the distance traveled by the host vehicle C is Y [m]. If there is, the lateral position movement amount determination flag is turned ON at time t14.

  That is, the determination unit 75b determines that the target TG is a pedestrian P that may collide with the host vehicle C at the time t14. Then, the determination unit 75b outputs information indicating the state of the lateral position movement amount determination flag to the count processing unit 73f.

  As described above, the movement condition 74d is such that the distance that the target moves in the direction intersecting the traveling direction of the host vehicle C is equal to or greater than the second threshold value, and It further includes that the traveling distance of the host vehicle C during the period of moving the distance of two or more thresholds is not more than the third threshold. Thereby, the determination unit 75b can more accurately determine the target actually moved in the horizontal direction as the pedestrian P, not the variation in the horizontal position of the target.

  When the target is a pedestrian, it moves 0.7 m or more in a direction intersecting the traveling direction of the host vehicle C in a relatively short time (for example, 1 second). On the other hand, when it is not a pedestrian (for example, in the case of a stationary object such as a stopped vehicle), the reflection point of one stopped vehicle is detected for each target detection process in a relatively long time (for example, 5 seconds). It is determined that the vehicle has moved more than 0.7 m despite the fact that it is a stop vehicle due to the variation in the so-called reflection points that become different reflection points (for example, the reflection point of the left brake lamp and the reflection point of the right brake lamp) There are things to do. Thus, even if it is erroneously determined that a target that is originally a stationary object has moved due to variation, the possibility of a pedestrian is reduced because the moving distance with respect to the passage of time is short.

  Therefore, when it corresponds to a relatively short time, in other words, when the traveling distance of the host vehicle C is Y [m] or less, the condition is satisfied, and the target is a target with high possibility of a pedestrian. If it is determined that it corresponds to a relatively long time, in other words, if it exceeds Y [m], the condition is not satisfied, and the target is determined to be a target with a low possibility of a pedestrian. .

  Thus, the movement condition 74d is a further condition that the traveling distance of the host vehicle C during the period in which the target moves a distance of the second threshold value or more in the direction intersecting the traveling direction of the host vehicle C is the third threshold value or less. Including. Thereby, the target actually moved in the horizontal direction can be more accurately determined as the pedestrian P.

  Next, the configuration of the peak extraction unit 73a and an example of the noise condition 74a will be described with reference to FIGS. FIG. 9 is a block diagram illustrating the peak extraction unit 73a according to the embodiment. FIG. 10 is an explanatory diagram illustrating an example of the noise condition 74a according to the embodiment.

  As shown in FIG. 9, the peak extraction unit 73a includes an extraction unit 75c and a noise determination unit 75d. The extraction unit 75 c is a processing unit that extracts the peak frequencies of the “UP section” and the “DN section” from the beat signal after the fast Fourier transform input from the FFT unit 72. The extraction unit 75c outputs the extracted peak frequency and peak to the azimuth calculation unit 73b and the noise determination unit 75d.

  The noise determination unit 75d is a processing unit that determines EMC noise using the characteristics of EMC noise. Specifically, the EMC noise is characterized in that the peak frequency on the “UP section” side and the peak frequency on the “DN section” side are substantially the same. On the other hand, when the host vehicle C is traveling, the peak on the “UP section” side of the roadside object is different from the peak on the “DN section” side.

  Further, the EMC noise is generated at a position relatively close to the radar apparatus 1 in the area where the radar apparatus 1 can detect the target, and the difference between the peak on the “UP section” side and the peak on the “DN section” side is generated. The signal intensity is also determined by the device that is the source of EMC noise.

  Therefore, the extraction unit 75c basically extracts a peak frequency whose peak exceeds the peak extraction threshold value. However, the extraction unit 75c, for example, in a frequency region having a frequency of 5 [BIN], even if the peak is equal to or less than the peak extraction threshold, the level of a predetermined signal strength determined by the device that generates the EMC noise described above. If there is, extract it.

  When the noise condition 74a stored in the storage unit 74 is satisfied for the peak frequency and peak on the “UP section” side and the “DN section” side input from the extraction unit 75c, the noise determination unit 75d performs EMC. Judged as noise.

  The noise condition 74a includes, for example, five conditions as shown in FIG. The first condition is that the peak frequency on the “UP section” side is 5 [BIN] or less. Here, the peak frequencies on the “UP section” side and the “DN section” side are values corresponding to the distance from the radar apparatus 1 to the target.

  BIN is a unit indicating a frequency on the frequency axis of the frequency spectrum graph. Therefore, the condition that the peak frequency on the “UP section” side is 5 [BIN] or less is that the peak frequency on the “UP section” side is a distance from 1.8 [m] to 8.5 [m] ahead of the radar apparatus 1. It means that the frequency corresponds to.

  The second condition is that the peak frequency on the “UP section” side is equal to the peak frequency on the “DN section” side. Note that the second condition does not necessarily need to completely match if the peak frequency on the “UP section” side and the peak frequency on the “DN section” side are substantially the same. For example, the second condition may be that the difference between the peak frequency on the “UP section” side and the peak frequency on the “DN section” side is 3 [BIN] or less.

  The third condition is that the absolute value of the value obtained by subtracting the peak on the “DN section” side from the peak on the “UP section” side is 2 [dB] or less. The fourth condition is that the peak on the “UP section” side is larger than −60 [dB] and not larger than −40 [dB]. The fifth condition is that the peak on the “DN section” side is larger than −60 [dB] and not larger than −40 [dB].

  The noise determining unit 75d determines that the noise condition 74a is satisfied for the peak frequency and peak on the “UP section” side and “DN section” side input from the extracting unit 75c, and determines that the noise is EMC noise. Set the judgment flag to ON.

  Further, the noise determination unit 75d determines that the noise condition 74a is not satisfied for the peak frequency and peak on the “UP section” side and the “DN section” side input from the extraction unit 75c, and determines that the noise is EMC noise. Set the judgment flag to OFF. Then, the noise determination unit 75d outputs information indicating the state of the EMC noise determination flag to the count processing unit 73f together with the peak frequency and the peak extracted by the extraction unit 75c.

  As described above, the noise condition 74a includes that the peak frequency on the “UP section” side and the peak frequency on the “DN section” side are substantially the same. Thereby, the noise determination part 75d can determine EMC noise correctly.

  The noise condition 74a is such that the absolute value of the difference between the peak (signal intensity) on the “UP section” side and the peak (signal intensity) on the “DN section” side is a predetermined value (for example, 2 [dB]) or less. Including that. Thereby, the noise determination part 75d can determine EMC noise more correctly.

  Next, the configuration of the count processing unit 73f and an example of the counter operation condition 74b will be described with reference to FIGS. FIG. 11 is a block diagram illustrating a count processing unit 73f according to the embodiment. FIG. 12 is an explanatory diagram illustrating an example of the counter operation condition 74b according to the embodiment.

  As shown in FIG. 11, the count processing unit 73f includes an operation unit 75e and a survival counter 75f. The operation unit 75e instructs the survival counter 75f to add, subtract, or hold the count value 74c based on information input from the filter processing unit 73e, the target classification unit 73g, and the peak extraction unit 73a. Is a processing unit for outputting.

  The survival counter 75f is a counter that performs addition, subtraction, holding, and the like of a count value 74c indicating the high possibility that a target exists in accordance with an instruction input from the operation unit 75e. In the present embodiment, the lower limit value of the count value 74c is “0”, and the upper limit value of the count value 74c is “35”.

  The radar apparatus 1 sets a target whose count value is larger than “0” as a tracking target candidate as a pedestrian, and when the target count value 74c becomes “0”, the tracking target candidate as a pedestrian. Exclude from

  When the information indicating that the extrapolation flag is OFF is input from the filter processing unit 73e, the operation unit 75e outputs a command for adding the count value 74c to the survival counter 75f. The survival counter 75f adds the count value 74c of the survival counter 75f and stores it in the storage unit 74 in accordance with a command from the operation unit 75e.

  In addition, when the information indicating that the extrapolation flag is ON is input from the filter processing unit 73e, the operation unit 75e receives information input from the target classification unit 73g and the peak extraction unit 73a, and counter operation conditions. 74b, the operation amount of the count value 74c is determined.

  For example, when the extrapolation flag is ON and the counter operation condition 74b is not satisfied, the operation unit 75e outputs a command for subtracting a predetermined value from the count value 74c to the survival counter 75f. When the extrapolation flag is ON and the counter operation condition 74b is satisfied, the operation unit 75e outputs a command to keep the current value as it is without subtracting a predetermined value from the count value 74c to the survival counter 75f. To do.

  The counter operation condition 74b includes, for example, seven conditions as shown in FIG. The first condition is that information indicating that the state of the previous lateral position movement amount determination flag was ON was input from the target classification unit 73g. As a result, the operation unit 75e can prohibit the process of holding the count value 74c for unnecessary targets such as objects below.

  The second condition is that information indicating that the current state of the EMC noise determination flag is OFF is input from the peak extraction unit 73a. As a result, the operation unit 75e can prohibit the process of holding the target count value 74c during the generation of EMC noise. That is, it is possible to prevent EMC noise from being erroneously detected as a pedestrian target and operating the survival counter.

  The third condition is that the peak frequency on the “UP section” side input from the peak extraction unit 73a is 5 [BIN] or less. As described above, this means that the peak frequency on the “UP section” side is a frequency corresponding to an area from 1.8 [m] to 8.5 [m] ahead of the radar apparatus 1. Such an area is an area where the peak extraction threshold is increased for EMC noise countermeasures.

  Accordingly, the operation unit 75e performs an operation of holding the count value 74c for a target detected in an area where a target that does not actually exist can be erroneously detected due to EMC noise. Therefore, the radar apparatus 1 can continuously detect the pedestrian P existing in the region where the peak extraction threshold is increased for EMC noise countermeasures.

  The fourth condition is that the peak on the “UP section” side is larger than −60 [dB]. Such a value of −60 [dB] is not EMC noise but signal intensity of floor noise generated by the radar apparatus 1 itself. Accordingly, the operation unit 75e performs an operation of holding the count value 74c for a target whose peak on the “UP section” side is higher than the signal intensity of the floor noise. Therefore, the operation unit 75e does not need to perform unnecessary processing on a target whose signal intensity is lower than the floor noise.

  The fifth condition is that the absolute value of the value obtained by subtracting the peak frequency on the predicted “UP section” side from the peak frequency on the “UP section” side is 3 [BIN] or less. The peak frequency on the “UP section” side in the fifth condition is the peak frequency on the “UP section” side extracted last time, and the peak frequency on the predicted “UP section” side is the “UP section” extracted last time. This is the peak frequency on the “UP section” side predicted from the peak frequency on the side.

  That is, the fifth condition is that the distance between the position corresponding to the peak frequency on the “UP section” side extracted last time and the position corresponding to the peak frequency on the “UP section” side predicted this time is the fourth threshold value. An example is 1.1 [m] or less.

  The sixth condition is that the absolute value of the value obtained by subtracting the peak frequency on the predicted “DN section” side from the peak frequency on the “DN section” side is 3 [BIN] or less. The peak frequency on the “DN section” side in the sixth condition is the peak frequency on the “DN section” side extracted last time, and the peak frequency on the predicted “DN section” side is the “DN section” extracted last time. This is the peak frequency on the “DN section” side predicted from the peak frequency on the side.

  That is, the sixth condition is that the distance between the position corresponding to the previously extracted peak frequency on the “DN section” side and the position corresponding to the peak frequency on the “DN section” side predicted this time is the fourth threshold value. An example is 1.1 [m] or less. The seventh condition is that the absolute value of the value obtained by subtracting the peak on the “DN section” side from the peak on the “UP section” side is 5 [dB] or less.

  As a result, the operation unit 75e counts the target value for which the distance from the target predicted this time based on the processing result of the previous detection process to the target detected in the previous detection process is equal to or less than the fourth threshold value. The operation of holding 74c is performed.

  As described above, the operation unit 75e performs an operation of holding the count value 74c by limiting to a target having a high possibility of the pedestrian P detected last time. Thereby, the radar apparatus 1 can continuously detect a target with a high possibility of the pedestrian P.

  When the extrapolation flag is ON and all of these seven conditions are satisfied, the operation unit 75e outputs a command for causing the survival counter 75f to hold the count value 74c. In addition, when any of these seven conditions is not satisfied, the operation unit 75e outputs a command for subtracting a predetermined value from the count value 74c to the survival counter 75f.

  The survival counter 75f holds the count value 74c in the storage unit 74 without subtraction when a command for holding the count value 74c is input from the operation unit 75e. The survival counter 75f subtracts the predetermined value from the count value 74c and stores it in the storage unit 74 when a command for subtracting the predetermined value from the count value 74c is input from the operation unit 75e.

  As described above, the count value 74c is referred to by the unnecessary target determining unit 73h and used for the unnecessary target determining process. For example, the unnecessary target determination unit 73h determines a target having a count value 74c larger than “0” as a target that is necessary for system control of a pedestrian or the like. Further, the unnecessary target determination unit 73h determines that the target whose count value 74c is “0” is an unnecessary target in terms of system control.

  That is, the count processing unit 73f functions as a candidate setting unit that sets a target as a tracking target candidate as a pedestrian by adding a predetermined value to the target count value 74c.

  The count processing unit 73 f functions as a candidate setting unit that excludes the target from the tracking target candidates as pedestrians by setting the target count value 74 c to “0”.

  Furthermore, the count processing unit 73f holds the target that has been subjected to extrapolation processing without subtracting the count value 74c of the target when it is determined that the target is a pedestrian in the previous determination processing. By doing so, it functions as a candidate setting unit that continuously sets the target as a tracking target candidate as a pedestrian without excluding it from the tracking target candidate as a pedestrian.

  Next, a processing procedure executed by the data processing unit 73 of the radar apparatus 1 according to the present embodiment will be described with reference to FIGS. 13, 14, 15, and 16. FIG. 13 is a flowchart illustrating main processing executed by the data processing unit 73 of the radar apparatus 1 according to the embodiment.

  FIG. 14 is a flowchart showing peak extraction processing during main processing according to the embodiment. FIG. 15 is a flowchart showing the target classification process in the main process according to the embodiment. FIG. 16 is a flowchart showing the counting process during the main process according to the embodiment.

  As shown in FIG. 13, first, the peak extraction unit 73a performs a peak extraction process based on the beat signal after the fast Fourier transform process input from the FFT unit 72 and the noise condition 74a (step S101). The details of the peak extraction process will be described later with reference to FIG. Subsequently, the azimuth calculation unit 73b performs an azimuth calculation process based on the processing result of the peak extraction process (step S102).

  Thereafter, the pairing unit 73c performs a pairing process based on the processing result of the azimuth calculation process (step S103). Subsequently, the continuity determination unit 73d performs a continuity determination process based on the processing result of the pairing process (step S104). Thereafter, the filter processing unit 73e performs a filter process based on the processing result of the continuity determination process (step S105).

  At this time, when the extrapolation process for predicting the virtual position of the target that is no longer detected in the current detection process is performed, the filter processing unit 73e turns on the extrapolation flag of the target. Further, when the extrapolation process is not performed, the filter processing unit 73e turns off the extrapolation flag of the target. Then, the filter processing unit 73e outputs information indicating the state of the extrapolation flag to the count processing unit 73f.

  Subsequently, the count processing unit 73f performs count processing based on the processing result of the current peak extraction processing, the processing result of the current filtering processing, the processing result of the target classification processing before this time, and the counter operation condition 74b ( Step S106). Details of the counting process will be described later with reference to FIG.

  Subsequently, the target classification unit 73g performs target classification processing based on the processing result of the filter processing (step S107). Details of the target classification process will be described later with reference to FIG. Thereafter, the unnecessary target determination unit 73h performs an unnecessary target determination process based on the processing result of the target classification process (step S108). Then, the combination processing unit 73i performs a combination process based on the processing result of the unnecessary target determination process (step S109).

  Then, the output target selection unit 73j performs output target selection processing based on the processing result of the combination processing (step S110), and outputs the target information of the target selected as the output target to the external device. The process ends.

  Next, the peak extraction process will be described with reference to FIG. The UP peak and DN peak shown in FIG. 14 are the peak frequencies and peaks on the “UP section” side and the “UN section” side. The peak extraction unit 73a repeatedly executes the process shown in FIG. 14 at a predetermined period.

  Specifically, as shown in FIG. 14, the peak extraction unit 73a first extracts the UP peak / DN peak (step S201), and then the noise condition 74a is determined for the extracted UP peak / DN peak. It is determined whether or not it is established (step S202).

  If the peak extraction unit 73a determines that the noise condition 74a is satisfied (step S202, Yes), the peak extraction unit 73a turns on the EMC noise determination flag (step S203), and moves the process to step S204.

  If the peak extraction unit 73a determines that the noise condition 74a is not satisfied (No at Step S202), the peak extraction unit 73a turns off the EMC noise determination flag (Step S206), and moves the process to Step S204.

  In step S204, the peak extraction unit 73a outputs information indicating the state of the EMC noise determination flag to the count processing unit 73f, and outputs the UP peak / DN peak to the azimuth calculation unit 73b and the count processing unit 73f (step S205). ), The process is terminated.

  Next, the target classification process will be described with reference to FIG. Note that the target information described in step S301 in FIG. 15 is information including an estimated angle, a distance, a relative speed, and a target position of each target subjected to the filtering process. The target classification unit 73g repeatedly executes the process shown in FIG. 15 at a predetermined period.

  Specifically, as shown in FIG. 15, the target classification unit 73g first determines whether or not the target information of each target is input from the filter processing unit 73e (step S301). If the target classification unit 73g determines that the target information is not input (No in step S301), the process ends.

  When the target classification unit 73g determines that the target information has been input (Yes in step S301), the target classification unit categorizes the target into a moving object and a stationary object based on the input target information. Classification is performed (step S302). Subsequently, the target classification unit 73g outputs the result of the target classification to the unnecessary target determination unit 73h (Step S303), and adds the target position included in the target information to the target history information 74e (Step S303). S304).

  Thereafter, the target classification unit 73g determines whether the target history information 74e includes a target position added before this time for the target for which target information has been input this time (step S305). . When the target classification unit 73g determines that there is no target position added before this time (No in step S305), the target classification unit 73g sets the first (new) target position for the target target ( Step S313), the process is terminated.

  Further, when the target classification unit 73g determines that there is a target position added before this time (step S305, Yes), the current target position and the target position added before this time Based on the above, it is determined whether or not a movement condition 74d for temporary determination of the lateral position movement amount is satisfied (step S306).

  Then, when the target classification unit 73g determines that the movement condition 74d for lateral position movement amount temporary determination is satisfied (step S306, Yes), the target position classification unit 73g turns on the horizontal position movement amount temporary determination flag (step S307). The process moves to step S308.

  When the target classification unit 73g determines that the movement condition 74d for the horizontal position movement amount temporary determination is not satisfied (step S306, No), it turns off the horizontal position movement amount temporary determination flag (step S311). The process moves to step S308.

  In step S308, the target classification unit 73g determines whether or not the movement condition 74d for determining the horizontal position movement amount is satisfied. When the target classification unit 73g determines that the moving condition 74d for determining the horizontal position movement amount is satisfied (step S308, Yes), the target classification unit 73g turns on the horizontal position movement amount determination flag (step S309) and performs the process. Move to step S310.

  When the target classification unit 73g determines that the moving condition 74d for determining the horizontal position movement amount is not satisfied (No in step S308), the target classification unit 73g turns off the horizontal position movement amount determination flag (step S312) and performs the process. Move to step S310.

  In step S310, the target classification unit 73g outputs information indicating the state of the lateral position movement amount determination flag to the count processing unit 73f, and ends the process.

  Next, the counting process will be described with reference to FIG. The count processing unit 73f repeatedly executes the process shown in FIG. 16 at a predetermined cycle. Specifically, as illustrated in FIG. 16, the count processing unit 73f first determines whether or not there is input of information indicating the state of the extrapolation flag from the filter processing unit 73e (step S401).

  When the count processing unit 73f determines that there is no input of information indicating the state of the extrapolation flag (No in step S401), the count processing unit 73f ends the process. If the count processing unit 73f determines that there is input of information indicating the state of the extrapolation flag (Yes in step S401), the count processing unit 73f determines whether the state of the extrapolation flag is OFF (step S402).

  If the count processing unit 73f determines that the state of the extrapolation flag is OFF (Yes in step S402), the count processing unit 73f causes the survival counter 75f to add a predetermined value to the count value 74c (step S403) and ends the process. To do.

  When the count processing unit 73f determines that the extrapolation flag is not OFF, that is, the extrapolation flag is ON (No in step S402), the count processing unit 73f determines whether the counter operation condition 74b is satisfied. Is performed (step S404).

  If the count processing unit 73f determines that the counter operation condition 74b is satisfied (step S404, Yes), the count processing unit 73f holds the count value 74c in the survival counter 75f (step S405), and ends the process.

  Further, when it is determined that the counter operation condition 74b is not satisfied (No at Step S404), the count processing unit 73f causes the survival counter 75f to subtract a predetermined value from the count value 74c (Step S406), and ends the process.

  Next, an example of a processing result by the radar apparatus 1 according to the embodiment will be described with reference to FIG. FIG. 17 is an explanatory diagram illustrating an example of a processing result obtained by the radar apparatus 1 according to the embodiment. In addition, the horizontal axis shown in FIG. 17 has shown time. The unit of time is the count number [CNT] of the internal clock.

  Also, the left vertical axis shown in FIG. 17 indicates the count value 74c of the survival counter 75f. The vertical axis on the right side shown in FIG. 17 indicates the state of the extrapolation flag and the distance [m] from the radar apparatus 1 to the vertical position of the target.

  Moreover, the graph shown by the triangular legend in FIG. 17 shows the transition of the count value 74c. The graph shown by the legend of x in FIG. 17 shows the transition of the vertical position of the target. A graph indicated by a square legend in FIG. 17 indicates the transition of the state of the extrapolation flag.

  As shown in FIG. 17, in the processing result of the target detection processing according to the embodiment, for example, the target is a radar device as the host vehicle C travels from time 430 [CNT] to 440 [CNT]. Gradually approaching one. Although not shown here, the target is a pedestrian P who is moving in a direction crossing the traveling direction of the host vehicle C.

  In the processing result shown in FIG. 17, the extrapolation flag is once turned ON at time 434 [CNT]. For this reason, the radar apparatus 1 subtracts the count value 74c when the time is 434 [CNT].

  Thereafter, in the processing result shown in FIG. 17, the extrapolation flag is turned OFF when the time is 435 [CNT], and the extrapolation flag is ON from the subsequent time 436 [CNT] to 440 [CNT]. In this case, if the distance to the target is shortened in a general radar device, the count value 74c is continuously subtracted from the time 436 [CNT].

  On the other hand, as shown in FIG. 17, the radar apparatus 1 according to the embodiment has a counter in the period from 435 [CNT] to 438 [CNT] even when the distance to the target is shortened. When the operation condition 74b is satisfied, the count value 74c can be held.

  Thereby, according to the radar apparatus 1 which concerns on embodiment, the target with the high possibility of the pedestrian P can be detected continuously. The reason why the count value 74c is subtracted at the time points 439 [CNT] and 440 [CNT] in FIG. 17 is that the target moves to the rear of the host vehicle C as the host vehicle C travels. is there.

  In the above-described embodiment, the radar apparatus 1 sets a high peak extraction threshold in a frequency region of 5 [BIN] or less. However, when the noise determination flag is OFF, the frequency region is higher than 5 [BIN]. It can also be lowered to the same level as the peak extraction threshold at. Thereby, the radar apparatus 1 can continue to more accurately detect the pedestrian P having a peak in a frequency region of 5 [BIN] or less when no EMC noise is generated.

  In the above-described embodiment, the count value 74c is held while the extrapolation flag is turned on. However, the radar apparatus 1 determines that the pedestrian P exists when the count value 74c is held. Then, information indicating that the extrapolation flag is OFF may be output to the external device.

  In the above-described embodiment, the radar apparatus 1 includes the noise determination unit 75d in the peak extraction unit 73a, and noise determination is performed based on the signal before the filter processing. May be provided after the filter processing unit 73e.

  Thereby, since the noise determination part 75d can perform noise determination based on the signal after a filter process, more accurate noise determination is attained.

  In the above-described embodiment, the count processing unit 73f adds, subtracts, or holds the target count value 74c, so that the target is set as a tracking target candidate as a pedestrian or excluded from the tracking target candidate. However, this is an example.

  The radar apparatus 1 according to the embodiment is configured to include a candidate setting unit that uses a target as a tracking target candidate as a pedestrian or excludes it from the tracking target candidate by an arbitrary procedure without using the count value 74c. Also good.

  For example, the radar apparatus 1 turns on the tracking target candidate flag when the target is detected by the detection process and is determined to be a pedestrian, and when the target is an extrapolation process target, The configuration may include a candidate setting unit that turns off the flag.

  In such a configuration, the candidate setting unit continues to turn on without turning off the tracking target candidate flag when the target that has been subjected to extrapolation processing has been determined to be a pedestrian in the previous determination processing. To do. Thereby, the candidate setting unit can continue to set the target as a tracking target candidate as a pedestrian without excluding the target as a pedestrian somewhat from the candidate.

  In the above-described embodiment, the number of the transmission antennas 4 of the radar apparatus 1 is one and the number of the reception antennas 5 is four. However, this is an example, and a plurality of targets can be detected. Other numbers may be used.

  In the above-described embodiment, ESPRIT is given as an example of the direction-of-arrival estimation method used by the radar apparatus 1. However, 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.

  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.

DESCRIPTION OF SYMBOLS 1 Radar apparatus 2 Transmission part 4 Transmission antenna 5 Reception antenna 6 Reception part 7 Signal processing apparatus 71 Transmission / reception control part 72 FFT part 73 Data processing part 73a Peak extraction part 73b Direction calculating part 73c Pairing part 73d Continuity determination part 73e Filter process Unit 73f count processing unit 73g target classification unit 73h unnecessary target determination unit 73i combination processing unit 73j output target selection unit 74 storage unit 74a noise condition 74b counter operation condition 74c count value 74d movement condition 74e target history information 75a classification unit 75b determination unit 75c extraction unit 75d noise determination unit 75e operation unit 75f survival counter A detection area C own vehicle P pedestrian TG target

Claims (11)

A detection unit that repeatedly performs a detection process for detecting a target approaching the host vehicle;
An extrapolation unit that performs an extrapolation process for predicting the virtual position of the target that is no longer detected in the current detection process based on the result of the previous detection process;
A determination unit that performs a determination process for determining in time series whether or not there is a possibility of collision with the host vehicle with respect to the same target that is the target of the detection process or the extrapolation process;
If the target that has been subject to the extrapolation process has been determined to be the pedestrian in the previous determination process, the target is not excluded from the tracking target candidates as the pedestrian, and the target continues. A radar apparatus comprising: a candidate setting unit that sets a tracking target candidate as a pedestrian. The determination unit
The radar apparatus according to claim 1, wherein the target is determined to be the pedestrian when the target moves in a direction crossing a traveling direction of the host vehicle. The determination unit
When a predetermined movement condition is established, the target is determined to be a target that moves in a direction that intersects the traveling direction of the host vehicle,
The movement condition is as follows:
The radar apparatus according to claim 2, comprising a distance that the target moves in a direction that intersects with a traveling direction of the host vehicle. The movement condition is as follows:
The distance that the target has moved in the direction intersecting the traveling direction of the host vehicle is equal to or greater than a second threshold value that is higher than the first threshold value, and the target is moved in the direction intersecting the traveling direction of the host vehicle. The radar apparatus according to claim 3, further comprising a distance traveled by the host vehicle being a third threshold value or less during a period of traveling a distance of two threshold values or more. The candidate setting unit
Each time the target is detected by the detection process, a predetermined value is added to a count value indicating the high possibility that the target exists, and the virtual position of the target is predicted by the extrapolation process. A counter unit that subtracts a predetermined value from the count value of the target each time,
When the target that has been subjected to the extrapolation processing has been determined to be the pedestrian in the previous determination processing, the count value is retained in place of subtraction of the count value in the counter unit. The radar device according to any one of claims 1 to 4, further comprising: A transmission unit that transmits a transmission wave that repeats an ascending period in which the frequency rises and a descent period in which the frequency falls to the target; a reflected wave of the transmission wave that is reflected by the target; And a receiving unit for generating a beat signal of the reflected wave,
A peak that performs fast Fourier transform on the beat signal and extracts, as a peak frequency, a frequency at which the signal intensity of the beat signal has a maximum maximum value for each of the rising and falling sections of the beat signal after the fast Fourier transform. An extraction unit and a noise determination unit that determines whether the peak frequency is noise based on the peak frequency, and
The detector is
Detecting the target by pairing the peak frequency in the rising section and the peak frequency in the falling section,
The operation unit is
The target detected by pairing and detecting the peak frequency determined to be noise by the noise determination unit is excluded from the target that causes the counter unit to hold the count value instead of subtracting the count value. The radar apparatus according to claim 5. The noise determination unit
When a predetermined noise condition is satisfied, the peak frequency is determined as noise,
The noise condition is
The radar apparatus according to claim 6, comprising: a peak frequency in the rising section of the target being substantially the same as a peak frequency in the falling section. The noise condition is
The absolute value of the difference between the signal intensity at the peak frequency in the ascending section of the target and the signal intensity at the peak frequency in the descending section further includes a predetermined value or less. Item 8. The radar device according to Item 7. The operation unit is
The counter unit is configured to hold the count value instead of subtracting the count value for a target whose signal intensity at the peak frequency in the rising section is higher than the signal intensity of floor noise. Item 8. The radar device according to any one of Items 6 to 8. The operation unit is
For a target detected by the detection unit in an area where the target that does not actually exist can be erroneously detected by electromagnetic waves emitted from other electronic devices, instead of subtracting the count value from the counter unit, The radar apparatus according to claim 5, wherein holding is performed. A target detection method executed by a computer,
Repeatedly performing a detection process of detecting a target approaching the host vehicle;
Performing an extrapolation process for predicting the virtual position of the target that is no longer detected in the current detection process based on the result of the previous detection process;
A step of performing a determination process for determining, in a time series, whether the pedestrian is likely to collide with the own vehicle with respect to the same target that is the target of the detection process or the extrapolation process;
If the target that has been subject to the extrapolation process has been determined to be the pedestrian in the previous determination process, the target is not excluded from the tracking target candidates as the pedestrian, and the target continues. And a step of setting a tracking target candidate as a pedestrian.

Images