WO2022249552A1

WO2022249552A1 - Information processing device and information processing method

Info

Publication number: WO2022249552A1
Application number: PCT/JP2022/004216
Authority: WO
Inventors: 亮佑山田
Original assignee: ソニーセミコンダクタソリューションズ株式会社
Priority date: 2021-05-24
Filing date: 2022-02-03
Publication date: 2022-12-01
Also published as: JP2022180111A

Abstract

This feature relates to an information processing device and an information processing method that enables highly robust object recognition to be performed with high resolution. In the present invention, a Fourier transform process is executed on a reception signal from an antenna to calculate a first angle spectrum, a high-resolution algorithm is executed on the reception signal to calculate a second angle spectrum, and each of the first and second angle spectra is outputted.

Description

Information processing device and information processing method

The present technology relates to an information processing device and an information processing method, and more particularly to an information processing device and an information processing method that enable object recognition with high resolution and high robustness.

Patent Documents 1 and 2 disclose radar devices that detect a target with high resolution using a high resolution algorithm.

Japanese Patent No. 5114217 Japanese Patent No. 6687289

In a radar device that uses a high-resolution algorithm to detect targets with high resolution, the robustness of the high-resolution algorithm is low, so the accuracy of object recognition tends to decrease when noise is large.

This technology was created in view of this situation, and enables object recognition with high resolution and high robustness.

An information processing apparatus according to the present technology includes a first spectrum processing unit that performs Fourier transform processing on a signal received from an antenna to calculate a first angle spectrum, and a high-resolution algorithm for the received signal. and an output unit configured to output the first angle spectrum and the second angle spectrum.

In the information processing method of the present technology, the first spectrum processing unit of an information processing device having a first spectrum processing unit, a second spectrum processing unit, and an output unit performs the following on a signal received from an antenna: , performing Fourier transform processing to calculate a first angle spectrum, the second spectrum processing unit performing a high-resolution algorithm on the received signal to calculate a second angle spectrum, In the information processing method, an output unit outputs the first angle spectrum and the second angle spectrum, respectively.

In the information processing device and information processing method of the present technology, Fourier transform processing is performed on a signal received from an antenna to calculate a first angle spectrum, and a high resolution algorithm is performed on the received signal. Then, a second angular spectrum is calculated, and the first angular spectrum and the second angular spectrum are respectively output.

1 is a diagram illustrating a configuration of a first embodiment of a radar device to which the present technology is applied; FIG. It is a figure showing the flow of the processing which the radar processing part of the radar installation of a 1st embodiment performs. It is the figure which illustrated the flow of the process of the 1st form of a reliability setting part. It is the figure which illustrated the flow of the process of the 2nd form of a reliability setting part. FIG. 3 is a diagram illustrating a configuration of a second embodiment of a radar device to which the present technology is applied; It is a figure showing the flow of the process which the radar processing part of the radar apparatus of 2nd Embodiment implements.

Embodiments of the present technology will be described below with reference to the drawings.

<First Embodiment of Radar Device>
FIG. 1 is a diagram illustrating the configuration of a first embodiment of a radar device to which the present technology is applied.

The radar device 1 in FIG. 1 is a device that detects the distance, direction, etc. of an object (target) existing in space by emitting radio waves and capturing the reflected waves. The radar device 1 is, for example, an FMCW (Frequency Modulated Continuous Wave) radar device that uses millimeter waves (30 GHz to 300 GHz in frequency) as radio waves. However, the present technology can also be applied to radar devices other than the FMCW system.

The radar device 1 has a transmitting antenna 11 , a receiving antenna 12 , an RF (Radio Frequency) front end section 13 , a radar processing section 14 , and a detection processing/tracking recognition section 15 .

The transmission antenna 11 radiates the transmission signal supplied from the RF front end section 13 into the air as radio waves (transmission waves).

The receiving antenna 12 receives radio waves (also referred to as received waves, reflected waves, or incoming waves) that arrive after being radiated from the transmitting antenna 11 and reflected by an object (target). A received wave received by the receiving antenna 12 is supplied to the RF front end section 13 as a received signal.

The receiving antenna 12 is composed of, for example, a plurality of receiving antennas (array antennas) arranged linearly, and the plurality of receiving antennas are shown as one receiving antenna 12 in FIG. When specifying a plurality of receive antennas, they are represented as receive antennas 12-1 to 12-N (N is the number of receive antennas). The receiving antennas 12 are not limited to linear array antennas, and may be planar (two-dimensional) array antennas.

The RF front end unit 13 generates a transmission signal and supplies it to the transmission antenna 11 and supplies an IF (Intermediate Frequency) signal corresponding to the signal received from the reception antenna 12 to the radar processing unit 14 .

The radar processing unit 14 calculates distance spectrum, angle spectrum, speed spectrum, etc. based on the IF signal from the RF front end unit 13 . The distance spectrum is the distance ( object distance). The angle spectrum is information specifying the direction of the position where the object exists (the direction of the object) in the scanning range of the radar device 1 . Regarding the direction of the object, for example, when the position of the radar device 1 (receiving antenna 12) is set as a reference position and a predetermined direction viewed from the reference position is set as a reference direction (for example, the center direction of the scanning range of the radar device 1), It is expressed by the angle of the direction from the reference position to the position where the object exists with respect to the reference direction. The direction of an object is also called the angle of the object. In the following description, directions or angles refer to directions or angles with respect to a predetermined reference direction viewed from a predetermined reference position. A velocity spectrum is information specifying the speed of movement of a moving object. Information such as the distance spectrum and angle spectrum obtained by the radar processing unit 14 is supplied to the detection processing/tracking recognition unit 15 .

The detection processing/tracking recognition unit 15 detects an object of interest (target: target) based on the information from the radar processing unit 14, and detects while specifying the distance and direction of the target with respect to the radar device 1. track the target. Note that the target may be a moving object, a predetermined type of object, or the like. Target information such as the distance and direction of the tracked target and tracking information (moving trajectory, etc.) are supplied to a processing unit (not shown) that performs processing such as imaging.

Note that the detection processing/tracking recognition unit 15 may be an arbitrary processing unit according to the application of the radar device 1, and detailed description thereof will be omitted. However, processing related to the present technology will be described later.

(Configuration and processing of RF front end unit 13)
The RF front end section 13 has a chirp signal generation section 31 ,

amplification sections

32 and 33 , a mixing section 34 , an LPF (low pass filter) section 35 and an A/D conversion section 36 .

The chirp signal generation unit 31 generates a chirp signal by frequency-modulating the sine wave signal, and supplies the chirp signal to the amplification unit 32 and the mixing unit 34 . A chirp signal is, for example, a signal whose frequency is continuously (linearly) changed (swept) from a predetermined minimum frequency to a predetermined maximum frequency at predetermined intervals.

The amplifier 32 amplifies the chirp signal from the chirp signal generator 31 and supplies it to the transmission antenna 11 .

The amplification section 33 amplifies the received signal from the receiving antenna 12 and supplies it to the mixing section 34 .

The mixing unit 34 generates an IF signal by mixing the chirp signal from the chirp signal generation unit 31 and the received signal from the amplification unit 33 . The IF signal is a beat signal having a difference frequency (beat frequency) that is the difference between the frequency of the received signal and the frequency of the chirp signal. The IF signal generated by the mixing section 34 is supplied to the LPF section 35 .

The LPF section 35 removes high-frequency components such as noise from the IF signal from the mixing section 34 and supplies it to the A/D conversion section 36 .

The A/D conversion section 36 samples the value of the IF signal from the LPF section 35 at a predetermined sampling period, and converts the sampled value from an analog value to a digital value. This converts the IF signal from an analog signal to a digital signal. The IF signal converted into a digital signal is supplied to the radar processing section 14 .

Note that IF signals for N channels corresponding to the plurality of receiving antennas 12-1 to 12-N in the receiving antenna 12 are supplied from the RF front end unit 13 to the radar processing unit 14. FIG. The RF front end unit 13 has an amplifier unit 33, a mixing unit 34, an LPF unit 35, and an A/D conversion unit 36 corresponding to each of the receiving antennas 12-1 to 12-N for N channels. However, any one or more of these processing units 33 to 36 perform processing for a plurality of channels by time division processing, so that the RF front end unit 13 can process the processing units 33 to 36 for N channels. You may not have it.

(Configuration and processing of radar processing unit 14)
The radar processing unit 14 is a processing unit configured by a DSP (Digital Signal Processor), and by executing a program, an FFT (Fast Fourier Transform) unit 51, an FFT unit 52, a high resolution algorithm processing unit 53, and a reliability A setting unit 54 is constructed.

The FFT unit 51 performs distance FFT and speed FFT processing on the IF signal from the A/D conversion unit 36 of the RF front end unit 13 .

Distance FFT converts the IF signal from the A/D converter 36 from time domain expression (expression with a function with time t as a variable) to frequency domain expression (expression with a function with frequency as a variable). FFT (Fast Fourier Transform) for frequency conversion. Distance FFT is performed on the IF signals of each channel corresponding to each of the receiving antennas 12-1 through 12-N. As a result, a spectrum (spectrum signal) showing high intensity at a frequency corresponding to the distance of the object (target) present in the entire scanning range of the radar device 1 is obtained. Since the frequency and the distance of an object have a certain relationship, the spectrum for the frequency (frequency spectrum) obtained by the distance FFT is the distance of the object corresponding to the frequency (the distance from the radar device 1 where the object can exist). , hereinafter simply referred to as distance). In the following description, the term "distance spectrum" means the spectrum for distance.

The velocity FFT is an FFT that performs frequency conversion from the time domain representation to the frequency domain representation for the component signal in which the data for the same distance are arranged in chronological order in the distance spectrum data obtained by the distance FFT. For example, M cycles ( The IF signal for chirp frame) is regarded as one set of IF signal. Since the distance FFT is performed for each IF signal of one chirp frame, if the distance FFT is performed for one set of IF signal, the data of the distance spectrum for M chirp frames is one set of distance spectrum. obtained as data for In the velocity FFT, frequency conversion is performed by FFT on the component signal (temporal component signal of the distance spectrum) in which M pieces of data for the same distance are arranged in chronological order in one set of distance spectrum data. will be As a result, a spectrum showing high intensity at frequencies corresponding to the moving speed of the object (relative speed between the radar device 1 and the object) is obtained. The velocity FFT is repeatedly performed each time one set of IF signals is supplied from the A/D conversion section 36 to the FFT section 51 . The velocity FFT may be performed only on the temporal component signal of the range spectrum for distances where the range spectrum determines that an object exists, or it may be performed on the temporal component signal of the range spectrum for distances over the entire range of the range spectrum. It may be performed on the component signals. In addition, since the frequency in the frequency domain when the frequency is converted by the velocity FFT and the moving speed of the object have a certain relationship, the spectrum for the frequency obtained by the velocity FFT (frequency spectrum) is the moving speed of the object corresponding to the frequency (velocity at which an object can move, hereinafter simply referred to as velocity). In the following description, the term "velocity spectrum" means a spectrum with respect to velocity. The moving speed of the object is not limited to the case where it is detected by the speed FFT, and may be the case where the speed FFT is not performed.

The FFT unit 51 converts information about the distance and moving speed of an object existing in the scanning range, such as the distance spectrum obtained by the distance FFT and the speed spectrum obtained by the speed FFT, into the FFT unit 52 as needed, and the high resolution It is supplied to the algorithm processing unit 53 , the reliability setting unit 54 , and the detection processing/tracking recognition unit 15 .

The FFT unit 52 (first spectrum processing unit) acquires distance spectrum and speed spectrum data calculated by the distance FFT and speed FFT (distance/speed FFT) of the FFT unit 51, and performs angle FFT processing (angular direction estimation process). The distance spectrum and velocity spectrum data obtained by the distance/velocity FFT are referred to as distance/velocity spectrum data. The FFT unit 52 (and the high-resolution algorithm processing unit 53) may not consider velocity, and in that case the distance/velocity spectrum corresponds to the distance spectrum.

The angle FFT is an FFT using the distance/velocity spectrum data of each channel obtained by the distance/velocity FFT for the IF signal of each channel corresponding to the plurality of receiving antennas 12-1 to 12-N. Specifically, in the angle FFT, in the distance/velocity spectrum data of each channel, N pieces (N channels) of data for the same distance and the same speed are obtained from the corresponding receiving antennas 12-1 to 12- The component signals (spatial component signals of the range spectrum) spatially arranged as values at the positions of N are subjected to frequency conversion from the spatial domain representation to the frequency domain representation by FFT. As a result, a spectrum showing high intensity at a frequency corresponding to the angle at which the object exists (the angle formed by the direction of the center of the scanning range of the radar device 1 and the direction of the object) is obtained. The angle FFT may be performed only on the spatial component signal of the range-velocity spectrum for the distances and velocities at which the range-velocity spectrum determines that an object is present, or it may be performed over the entire range of the range-velocity spectrum. may be performed on the spatial component signals of the distance-velocity spectrum for the distances and velocities of . Since the frequency in the frequency domain when the frequency is converted by the angle FFT and the angle of the object have a certain relationship, the spectrum for the frequency in the frequency domain when the frequency is converted by the angle FFT (frequency spectrum) It can be regarded as a spectrum with respect to the angle of the corresponding object (the angle (direction) with respect to the radar device 1 (receiving antenna 12) at the position where the object may exist, hereinafter simply referred to as the angle). In the following description, the term "angle spectrum" means a spectrum with respect to an angle.

The FFT unit 52 supplies information about the angles of objects existing in the scanning range, such as the angle spectrum obtained by the angle FFT, to the reliability setting unit 54 and the detection processing/tracking recognition unit 15 as necessary.

The high-resolution algorithm processing unit 53 (second spectrum processing unit) performs high-resolution processing based on the IF signal data of each channel from the A/D conversion unit 36 or the distance/velocity spectrum data from the FFT unit 51. Using an algorithm, angle estimation processing (direction-of-arrival estimation) with higher resolution than the FFT unit 52 is performed.

Here, the angular spectrum obtained by the FFT unit 52 is the result of estimating the direction of arrival of the incoming wave (received wave) received by the receiving antenna 12 using a beamformer method based on Fourier transform as a method of estimating the direction of arrival. be. The beamformer method has a lower resolution than the direction-of-arrival estimation method using a high-resolution algorithm, but the amount of computation is small, so the load and time required for computation processing are small. Therefore, the calculation of the angle spectrum by the FFT unit 52 is performed in a short time.

The high-resolution algorithm used in the high-resolution algorithm processing unit 53 has a larger amount of calculation than the beamformer method, so the load and time required for calculation processing are large, but the resolution is high. In this embodiment, a high-resolution algorithm refers to any direction-of-arrival estimation method that has higher resolution than the beamformer method. High-resolution algorithms include Capon method, CS method (compressed sensing), linear prediction method (LP: Linear Prediction), Pisarenko method, MUSIC method (MUltiple SIgnal Classication), ESPRIT method (Estimation of Signal Parameters via Rotational Invariance Techniques), Deterministic Maximum Likelihood, Weighted Subspace Fitting, Root-MUSIC, etc. are well known. The high-resolution algorithm processing unit 53 may use any direction-of-arrival estimation method among these well-known high-resolution algorithms having higher resolution than the beamformer method.

For example, when using the MUSIC method as the high-resolution algorithm, the high-resolution algorithm processing unit 53 calculates the MUSIC spectrum (evaluation function) using the IF signal data of each channel and the steering matrix. The MUSIC spectrum (also referred to as MUSIC spectrum) corresponds to the angle spectrum generated by the angle FFT of the FFT unit 52, and exhibits high intensity for angles at which objects exist. The MUSIC spectrum has higher resolution than the angle spectrum obtained by angle FFT. Note that the component values in the column direction of the steering matrix are the amplitudes caused by the phase differences between the reception signals received by the reception antennas 12-1 to 12-N according to the arrival angles of the radio waves (received waves). represents vibration. In each column of the steering matrix, component values are arranged when the arrival angle of the received radio wave is changed by a predetermined angle.

The high-resolution algorithm processing unit 53 outputs the results of the angle estimation processing using the high-resolution algorithm (angle spectra for each distance and speed at predetermined intervals) to the reliability setting unit 54 or detection processing/tracking recognition as necessary. 15.

Based on the angle spectrum obtained by the FFT unit 52, the reliability setting unit 54 sets the reliability of the result of the angle estimation processing using the angle FFT of the FFT unit 52 and the high resolution algorithm of the high resolution algorithm processing unit 53. The reliability of the result of the angle estimation process used is calculated. The processing of the reliability setting unit 54 will be described later. The reliability setting unit 54 supplies the reliability of the result of the angle estimation processing using the angle FFT and the reliability of the result of the angle estimation processing using the high resolution algorithm to the detection processing/tracking recognition unit 15 .

<Processing Flow of Radar Processing Unit 14 of Radar Device 1>
FIG. 2 is a diagram showing the flow of processing performed by the radar processing unit 14 of the radar device 1. As shown in FIG.

In step S11, the FFT unit 51 of the radar processing unit 14, via the receiving antennas 12 (12-1 to 12-N) and the RF front end unit 13, corresponds to each of the receiving antennas 12-1 to 12-N. The digital value of the received signal (IF signal) of the channel is acquired at a predetermined sampling period.

In step S12, the FFT unit 51 performs distance FFT and velocity FFT using the IF signal data acquired in step S11. Thereby, the distance spectrum and velocity spectrum (distance/velocity spectrum) in each channel are calculated.

In step S13, the FFT unit 52 uses the distance/velocity spectrum data of each channel calculated in step S12 to calculate the angle spectrum by angle FFT, and performs angle estimation processing. The angle spectrum is calculated for each distance and speed at predetermined intervals within the distance range and (velocity range) of the distance/velocity spectrum. A result D1 in FIG. 2 is a diagram exemplifying the result of angle estimation processing performed by the FFT unit 52 using the angle FFT. In the result D1, the intensity of the angle spectrum calculated using the angle FFT for each distance from the radar device 1 at predetermined intervals within the scanning range of the radar device 1 is compared with the corresponding distance and angle position. It is shown in image density. According to this, a region with a higher image density (a region with a higher angular spectrum intensity) is represented as a region with a higher possibility that an object exists.

In step S14, the high-resolution algorithm processing unit 53 performs high-resolution angle estimation processing using the high-resolution algorithm. Result D2 in FIG. 2 is a diagram exemplifying the result of angle estimation processing performed by the high-resolution algorithm processing unit 53 using the high-resolution algorithm. In the result D2, within the scanning range of the radar device 1, for each distance from the radar device 1 at predetermined intervals, the position corresponding to the angle (the angle at which the object exists) estimated using the high-resolution algorithm. A cross is indicated.

In step S15, the detection processing/tracking recognition unit 15 combines the result D1 obtained by the angle estimation processing using the angle FFT in step S13 and the result D2 obtained by the angle estimation processing using the high-resolution algorithm in step S14. , Recognition of an object (including recognition of existence, size, distance, direction, etc. of the object) existing in the scanning range of the radar device 1 is performed based on the reliability (described later) for each of the results D1 and D2.

<Processing of the first form of the reliability setting unit 54>
FIG. 3 is a diagram exemplifying the flow of the first mode of processing by the reliability setting unit 54. As shown in FIG. In FIG. 3, steps S31 to S33 are performed by the FFT unit 52 in step S13 of FIG. 2, where the FFT unit 52 performs angle estimation processing using the angle FFT. 15 is performed before performing object recognition.

In step S31, the reliability setting unit 54 detects the noise level (noise floor) of an area where there are no objects and only noise, based on the result D1 in step S31. In result D1, for example, no object exists in region R1. The reliability setting unit 54 determines the intensity (power ) is detected (calculated) as the noise floor. In this specification, the value corresponding to the magnitude of the amplitude value (complex number) at each angle of the angle spectrum is referred to as the intensity (or power) of the angle spectrum. For example, the absolute value of the amplitude value at each angle of the angle spectrum, the square of the absolute value of the amplitude value (power spectrum), and the value obtained by dividing the square of the absolute value of the amplitude value by the angular resolution (angular interval) of the angle spectrum ( power spectral density) can be used as the intensity (or power) of the angular spectrum.

In step S32, the reliability setting unit 54 increases the reliability of the result D1 of the angle estimation processing performed by the FFT unit 52 using the angle FFT in step S13 of FIG. 2 as the noise floor detected in step S31 increases. set to a value. The reliability setting unit 54 sets the reliability of the result D1 to a lower value as the noise floor detected in step S31 is smaller.

In step S33, the reliability setting unit 54 determines that the higher the noise floor detected in step S31, the higher is the angle estimation processing result D2 performed by the high-resolution algorithm processing unit 53 using the high-resolution algorithm in step S14 of FIG. Set confidence to a low value. The reliability setting unit 54 sets the reliability of the result D2 to a higher value as the noise floor detected in step S31 is smaller.

In step S15, the detection processing/tracking recognition unit 15 generates the result D1 of the angle estimation processing using the angle FFT in step S13, the result D2 of the angle estimation processing using the high-resolution algorithm in step S14, steps S32 and Based on the reliability of each of the results D1 and D2 set in step S33, the distance, direction (angle), etc. of an object existing within the scanning range of the radar device 1 are recognized. For example, the detection processing/tracking recognition unit 15 uses the result D1 and the result D2, whichever has the higher reliability value, to recognize an object existing in the scanning range (existence, size, distance, direction, etc.). recognition). The detection processing/tracking recognition unit 15 weights the result D1 and the result D2 with respective reliability, and combines the results of both the result D1 and the result D2 to recognize the object existing in the scanning range. It may be the case.

According to this, when the noise floor detected by the reliability setting unit 54 in step S31 is larger than a predetermined level, the reliability of the result D1 is set to a value higher than the reliability of the result D2. can be done. If the noise is large and the result D2 of angle estimation processing using a low-robustness high-resolution algorithm is used for object recognition or the like, object recognition may be erroneous or impossible. Therefore, when the noise floor is larger than a predetermined level, the result D1 of angle estimation processing using a highly robust angle FFT (beamformer method) is used for object recognition, etc., thereby increasing the recognition accuracy. decrease in

Conversely, if the noise floor is below a predetermined level, the reliability of result D2 can be set to a higher value than the reliability of result D1. When the noise floor is lower than a predetermined level, object recognition can be performed appropriately using the result D2 of angle estimation processing using a low-robustness high-resolution algorithm. be done.

<Second Mode Processing of Reliability Setting Unit 54>
FIG. 4 is a diagram exemplifying the flow of the second mode of processing by the reliability setting unit 54. As shown in FIG. In FIG. 4, steps S51 to S53 are performed by the FFT unit 52 in step S13 of FIG. 2, where the FFT unit 52 performs angle estimation processing using the angle FFT, and after obtaining the result D1, in step S15, the detection processing/tracking recognition unit 15 is performed before performing object recognition.

In step S51, the reliability setting unit 54 detects the noise floor and the level of the peak region where the object is most likely to exist, based on the result D1 in step S31. In result D1 shown in FIG. 4, there is no object in region R1, for example. The reliability setting unit 54 determines the intensity (power ) is detected (calculated) as the noise floor. In the result D1 shown in FIG. 4, for example, the region R2 has the peak of the image density corresponding to the intensity of the angular spectrum, and is most likely to contain an object. The reliability setting unit 54 detects the maximum value of the intensity of the angle spectrum of the distance range and the angle range of the area R2 as the level of the peak area based on the angle spectrum corresponding to the distance range and the angle range of the area R2. Note that the minimum value or average value of the intensity of the angle spectrum in the distance range and the angle range of the region R2 may be detected as the level of the peak region. The reliability setting unit 54 calculates the level of the peak area with respect to the noise floor as a power ratio (intensity ratio) based on the detected noise floor and the level of the peak area. When the absolute value of the amplitude value at each angle of the angle spectrum is used as the intensity of the angle spectrum, the power ratio corresponds to the S/N ratio.

In step S52, the reliability setting unit 54 increases the reliability of the result D1 of the angle estimation processing performed by the FFT unit 52 using the angle FFT in step S13 of FIG. 2 as the power ratio calculated in step S51 decreases. set to a value. The reliability setting unit 54 sets the reliability of the result D1 to a lower value as the power ratio calculated in step S51 increases.

In step S53, the smaller the power ratio calculated in step S51, the higher the reliability setting unit 54 is. Set confidence to a low value. The reliability setting unit 54 sets the reliability of the result D2 to a higher value as the power cost calculated in step S31 increases.

In step S15, the detection processing/tracking recognition unit 15 generates the result D1 of the angle estimation processing using the angle FFT in step S13, the result D2 of the angle estimation processing using the high resolution algorithm in step S14, and the result D2 of the angle estimation processing using the high resolution algorithm in step S14. And based on the reliability of each of the results D1 and D2 set in step S33, the distance and direction (angle) of the object existing in the scanning range of the radar device 1 are recognized. For example, the detection processing/tracking recognition unit 15 uses the result D1 and the result D2, whichever has the higher reliability value, to recognize an object existing in the scanning range (existence, size, distance, direction, etc.). recognition). The detection processing/tracking recognition unit 15 weights the result D1 and the result D2 with respective reliability, and combines the results of both the result D1 and the result D2 to recognize the object existing in the scanning range. It may be the case.

According to this, when the power ratio calculated by the reliability setting unit 54 in step S51 is smaller than a predetermined level, the reliability of the result D1 is set to a value higher than the reliability of the result D2. can be done. If the noise is large and the power ratio is small, and the result D2 of angle estimation processing using a low-robustness high-resolution algorithm is used for object recognition or the like, object recognition may be erroneous or impossible. Therefore, when the power ratio is smaller than a predetermined level, the result D1 of angle estimation processing using a highly robust angle FFT (beamformer method) is used for object recognition, etc., thereby increasing the recognition accuracy. decrease in

Conversely, if the power ratio is greater than a predetermined level, the reliability of result D2 can be set to a value higher than the reliability of result D1. When the power ratio is greater than a predetermined level, object recognition can be performed appropriately using the result D2 of angle estimation processing using a low-robustness high-resolution algorithm. be done.

Note that the reliability setting unit 54 may set the reliability based on the distance of the target object (target). For example, when the distance to the target is less than or equal to a predetermined threshold, the reliability setting unit 54 determines that the result of the angle estimation process using the high-resolution algorithm is higher than the reliability of the result D1 of the angle estimation process using the angle FFT. Set the confidence of D2 to a high value. When the distance of the target is equal to or greater than a predetermined threshold, the reliability setting unit 54 selects the angle estimation using the angle FFT (beamformer method) rather than the reliability of the result D2 of the angle estimation processing using the high resolution algorithm. The reliability of the processing result D1 may be set to a high value.

Conventionally, either FFT or high-resolution algorithms are used for angle estimation processing in radar equipment. Angle estimation processing using FFT is characterized by high robustness but low resolution. In the case of angle estimation processing using a high-resolution algorithm, resolution is high, but robustness is low compared to FFT, and the amount of computation is enormous.

The present technology combines these different methods of angle estimation processing, and depending on the situation, uses the result of the appropriate method of angle estimation processing for processing such as object recognition. Therefore, it is possible to perform processing such as object recognition that makes the most of the advantages of each method.

<Second Embodiment of Radar Device>
FIG. 5 is a diagram illustrating the configuration of a radar device according to a second embodiment to which the present technology is applied. In addition, the same code|symbol is attached|subjected to the part which is common in the radar apparatus 1 of 1st Embodiment of FIG. 1, and the detailed description is abbreviate|omitted suitably.

The radar device 101 in FIG. 5 is common with the radar device 1 in FIG. .

The radar processing unit 14 of the radar device 101 of FIG. 5 is common to the radar processing unit 14 of the radar device 101 of FIG.

However, the radar processing unit 14 of the radar device 101 in FIG. 5 does not have the reliability setting unit 54 in FIG. 1, the FFT unit 121 is provided instead of the FFT unit 52 in FIG. 1 in that a high-resolution algorithm processing unit 122 is provided in place of the high-resolution algorithm processing unit 53 of FIG.

In the radar processing unit 14 of the radar device 101 in FIG. 5, the FFT unit 121 acquires distance spectrum data calculated by the distance FFT of the FFT unit 51, and performs angle FFT processing (angular direction estimation processing). Angular direction estimation processing using angle FFT (direction-of-arrival estimation using the beamformer method) is the same as that of the FFT unit 52 in the radar processing unit 14 of the radar device 1 in FIG. 1, so description thereof will be omitted.

The FFT unit 121 determines a high-resolution scanning range for angular direction estimation processing using a high-resolution algorithm, out of the scanning range of the radar device 1, as a result of the angular direction estimation processing using the angle FFT. For example, the high-resolution scan range may be the range of objects present near the center of the scan range or the range of moving objects, and the area of interest is determined as the high-resolution scan range. The FFT unit 121 supplies information indicating the determined high resolution scanning range (information indicating distance range and angle range) to the high resolution algorithm processing unit 122 .

The high-resolution algorithm processing unit 122, based on the data of the IF signal of each channel from the A/D conversion unit 36 or the data of the distance/velocity spectrum from the FFT unit 51, the scanning range of the radar device 1 High-resolution angle estimation processing is performed using a high-resolution algorithm only for the high-resolution scanning range indicated by the information supplied from the FFT unit 52 . The high-resolution algorithm processing unit 122 supplies the detection processing/tracking recognition unit 15 with the results of the angle estimation processing performed using the high-resolution algorithm for each distance and speed in the high-resolution scanning range.

<Processing Flow of Radar Processing Unit 14 of Radar Device 101>
FIG. 6 is a diagram showing the flow of processing performed by the radar processing unit 14 of the radar device 101. As shown in FIG.

In FIG. 6, step S81 is performed after the FFT unit 51 uses the data of the IF signal to calculate the distance spectrum and velocity spectrum in each channel by distance FFT and velocity FFT.

In step S81, the FFT unit 121 uses the distance/velocity spectrum data of each channel calculated by the FFT unit 51 to calculate the angle spectrum by angle FFT, and performs angle estimation processing. The angle spectrum is calculated for each distance and speed at predetermined intervals within the distance range and speed range of the distance/speed spectrum. A result D1 in FIG. 6 is a diagram exemplifying the result of angle estimation processing performed by the FFT unit 121 using the angle FFT. In the result D1, the intensity of the angle spectrum calculated using the angle FFT for each distance from the radar device 1 at predetermined intervals within the scanning range of the radar device 1 is compared with the corresponding distance and angle position. It is shown in image density. According to this, a region with a higher image density (a region with a higher angular spectrum intensity) is represented as a region with a higher possibility that an object exists.

Based on the result D1, the FFT unit 121 determines a range that meets predetermined conditions among the scanning range of the radar device 101 as a high-resolution scanning range. Assume, for example, that the FFT unit 121 determines the high-resolution scanning range P1 for the result D1 in FIG.

In step S82, the high-resolution algorithm processing unit 53 performs high-resolution angle estimation processing using the high-resolution algorithm only on the high-resolution scanning range P1 determined in step S81. Result D2 in FIG. 6 is a diagram exemplifying the result of angle estimation processing performed by the high-resolution algorithm processing unit 53 using the high-resolution algorithm. In the result D2, only the position (distance and angle) of the object existing in the high-resolution scanning range P1 determined in step S81 within the scanning range of the radar device 1 is detected.

In step S83, the detection processing/tracking recognition unit 15 combines the result D1 obtained by the angle estimation processing using the angle FFT (beamformer method) in step S81 and the angle estimation processing using the high-resolution algorithm in step S82. Based on the obtained result D2, an object existing in the scanning range is recognized (including recognition of existence, size, distance, direction, etc. of the object). At this time, the detection processing/tracking recognition unit 15 appropriately acquires the information of the high resolution scanning range P1 in the result D1 from the information of the high resolution scanning range P2 in the result D2.

According to the radar device 101, which is the second embodiment of the radar device to which the present technology is applied, the resolution of only an important range within the scanning range of the radar device 101 can be increased. Low resolution FFT (beamformer method) provides information even for unimportant ranges. As a result, the efficiency of arithmetic processing can be improved, and the load and time required for arithmetic processing can be reduced.

The present technology can also take the following configurations.
(1)
a first spectrum processing unit that performs Fourier transform processing on a signal received from an antenna to calculate a first angle spectrum;
a second spectrum processing unit that executes a high-resolution algorithm on the received signal to calculate a second angular spectrum;
An information processing apparatus comprising: an output unit that outputs the first angle spectrum and the second angle spectrum, respectively.
(2)
a reliability calculation unit that calculates a first reliability regarding the first angle spectrum and a second reliability regarding the second angle spectrum;
The above-described (1), further comprising: a recognition unit that performs object recognition by combining the first angle spectrum and the second angle spectrum based on the first reliability and the second reliability. information processing equipment.
(3)
The information according to (2), wherein the reliability calculation unit sets the first reliability to a higher value than the second reliability when noise in the first angular spectrum is greater than a predetermined level. processing equipment.
(4)
The information processing apparatus according to (2) or (3), wherein the reliability calculation unit sets the first reliability to a higher value as the noise of the first angle spectrum increases.
(5)
The information processing apparatus according to any one of (2) to (4), wherein the reliability calculation unit sets the second reliability to a lower value as the noise of the first angular spectrum increases.
(6)
The reliability calculation unit calculates the first reliability from the second The information processing apparatus according to (2) above, wherein the value is higher than the reliability of .
(7)
The reliability calculation unit sets the first reliability to a higher value as the ratio of the maximum value of the intensity of the first angle spectrum to the noise level of the first angle spectrum decreases. Or the information processing device according to (6).
(8)
The reliability calculation unit sets the second reliability to a lower value as the ratio of the maximum value of the intensity of the first angle spectrum to the noise level of the first angle spectrum decreases. , (6), or the information processing apparatus according to (7).
(9)
The information processing apparatus according to any one of (2) to (8), wherein the reliability calculation unit calculates the first reliability and the second reliability based on a distance of an object.
(10)
The information processing apparatus according to (9), wherein the reliability calculation unit sets the second reliability to a higher value than the first reliability when the distance of the object is equal to or less than a predetermined threshold.
(11)
The information according to (9) or (10), wherein the reliability calculation unit sets the first reliability to a higher value than the second reliability when the distance of the object is equal to or greater than a predetermined threshold. processing equipment.
(12)
The second spectrum processing unit calculates the first angle spectrum in a partial range with respect to the angle range for which the first spectrum processing unit calculates the first angle spectrum. The information processing device described.
(13)
the antenna comprises a plurality of antennas,
The information processing apparatus according to any one of (1) to (12), wherein the received signal is composed of received signals of a plurality of channels received by each of the plurality of antennas.
(14)
The first spectrum processing unit,
calculating the frequency spectrum of the plurality of channels for each of the received signals of the plurality of channels by performing the Fourier transform processing on each of the received signals of the plurality of channels; The information processing apparatus according to (13), wherein the first angle spectrum is calculated by performing the Fourier transform process on a component signal composed of component values.
(15)
The first spectrum processing unit and the second spectrum processing unit calculate the first angle spectrum and the second angle spectrum for each distance and speed at predetermined intervals from the antenna, respectively. ) to (14).
(16)
The spectrum processing unit of 1,
The information processing device according to any one of (1) to (15), wherein the first angle spectrum is calculated using a beamformer method.
(17)
The second spectrum processing unit,
As the high-resolution algorithm, the Capon method, the CS method, the linear prediction method, the Pisarenko method, the MUSIC method, and any one of the ESPRIT method for estimating the direction of arrival to calculate the second angle spectrum. The information processing apparatus according to any one of 1) to (16).
(18)
a first spectrum processing unit;
a second spectral processing unit;
The first spectrum processing unit of the information processing apparatus having an output unit performs Fourier transform processing on a signal received from an antenna to calculate a first angle spectrum,
The second spectrum processing unit calculates a second angle spectrum by executing a high-resolution algorithm on the received signal,
The information processing method, wherein the output unit outputs the first angle spectrum and the second angle spectrum, respectively.

1, 101 radar device, 11 transmitting antenna, 12 receiving antenna, 13 RF front end unit, 14 radar processing unit, 15 detection processing/tracking recognition unit, 51, 52, 121 FFT unit, 53, 122 high resolution algorithm processing unit, 54 Reliability setting part

Claims

a first spectrum processing unit that performs Fourier transform processing on a signal received from an antenna to calculate a first angle spectrum;
a second spectrum processing unit that executes a high-resolution algorithm on the received signal to calculate a second angular spectrum;
An information processing apparatus comprising: an output unit that outputs the first angle spectrum and the second angle spectrum, respectively.
a reliability calculation unit that calculates a first reliability regarding the first angle spectrum and a second reliability regarding the second angle spectrum;
The recognition unit according to claim 1, further comprising a recognition unit that performs object recognition by combining the first angle spectrum and the second angle spectrum based on the first reliability and the second reliability. Information processing equipment.
The information processing according to claim 2, wherein the reliability calculation unit sets the first reliability to a higher value than the second reliability when noise in the first angular spectrum is greater than a predetermined level. Device.
The information processing apparatus according to claim 2, wherein the reliability calculation unit sets the first reliability to a higher value as the noise of the first angle spectrum increases.
The information processing apparatus according to claim 2, wherein the reliability calculation unit sets the second reliability to a lower value as the noise of the first angle spectrum increases.
The reliability calculation unit calculates the first reliability from the second The information processing apparatus according to claim 2, wherein the value is higher than the reliability of .
3. The reliability calculation unit sets the first reliability to a higher value as the ratio of the maximum value of the intensity of the first angle spectrum to the noise level of the first angle spectrum is smaller. The information processing device described.
3. The reliability calculation unit sets the second reliability to a lower value as the ratio of the maximum value of the intensity of the first angle spectrum to the noise level of the first angle spectrum decreases. The information processing device described.
The information processing apparatus according to claim 2, wherein the reliability calculation unit calculates the first reliability and the second reliability based on a distance of an object.
The information processing apparatus according to claim 9, wherein the reliability calculation unit sets the second reliability to a higher value than the first reliability when the distance of the object is equal to or less than a predetermined threshold.
The information processing apparatus according to claim 9, wherein the reliability calculation unit sets the first reliability to a higher value than the second reliability when the distance of the object is equal to or greater than a predetermined threshold.
2. The second spectrum processing section according to claim 1, wherein the first spectrum processing section calculates the first angle spectrum within a partial range of the angle range for which the first spectrum processing section calculates the first angle spectrum. information processing equipment.
the antenna comprises a plurality of antennas,
The information processing apparatus according to claim 1, wherein the received signal comprises received signals of a plurality of channels received by each of the plurality of antennas.
The first spectrum processing unit,
calculating the frequency spectrum of the plurality of channels for each of the received signals of the plurality of channels by performing the Fourier transform processing on each of the received signals of the plurality of channels; The information processing apparatus according to claim 13, wherein the first angle spectrum is calculated by subjecting a component signal composed of component values to the Fourier transform process.
2. The first spectrum processing unit and the second spectrum processing unit calculate the first angle spectrum and the second angle spectrum for each distance and speed at predetermined intervals from the antenna, respectively. The information processing device according to .
The spectrum processing unit of 1,
The information processing apparatus according to claim 1, wherein the first angular spectrum is calculated using a beamformer method.
The second spectrum processing unit,
The second angular spectrum is calculated using any direction-of-arrival estimation method among the Capon method, CS method, linear prediction method, Pisarenko method, MUSIC method, and ESPRIT method as the high-resolution algorithm. 1. The information processing device according to 1.
a first spectrum processing unit;
a second spectral processing unit;
The first spectrum processing unit of the information processing apparatus having an output unit performs Fourier transform processing on a signal received from an antenna to calculate a first angle spectrum,
The second spectrum processing unit calculates a second angle spectrum by executing a high-resolution algorithm on the received signal,
The information processing method, wherein the output unit outputs the first angle spectrum and the second angle spectrum, respectively.