Optical processing device adaptive to parallel processing of multi-channel foresight radar signals

Abstract

An optical processing device adaptive to parallel processing of multi-channel foresight radar signals preprocesses high-flux radar signals in a simulation domain: high-complexity links in a radar imaging algorithm are moved to an optical domain for instantaneous parallel computing, and downsampling extraction is performed on data volume in the processing process; and the requirements of imaging on rear-end electric sampling and calculation are reduced, and radar foresight imaging with high real-time performance can be realized.

Description

Technical Field

[0001] The invention belongs to the field of optical computing technology, and in particular to an optical processing device adapted for parallel processing of multi-channel forward-looking radar signals. Background Art

[0002] Multi-channel forward-looking radar, with a large azimuth real aperture, can fundamentally solve the problem of no Doppler and left-right blur in the forward-looking imaging area, so that the radar has the ability to obtain information in all directions. Although the advantages are obvious, multi-channel forward-looking radar faces the processing bottleneck of high data volume and high computing volume: the MIMO system, non-scanning, and ultra-wideband characteristics of forward-looking imaging (resolution depends on the size of the physical aperture and bandwidth, and imaging quality depends on the MIMO scale and bandwidth) lead to huge data volume. Under the traditional electrical domain processing architecture, the processing delay is high, and the real-time advantage of multi-channel forward-looking radar is difficult to play; a large number of digital processing resources need to be piled up, and each additional receiving channel requires an additional set of ultra-high-speed ADC and processor; high-frequency signal transmission loss is high and stability is poor, and the synchronization accuracy between all channels needs to reach 20ps. As radar signals develop towards millimeter wave frequencies and ultra-wideband, the processing speed of electrical ADCs and processors will inevitably be unable to cope with it. Studies have shown that optical computing has obvious advantages over electrical architecture in multi-channel parallel processing speed, transmission, bandwidth, and synchronization stability, and has the potential to adapt to multi-channel forward-looking imaging.

[0003] However, although some studies have attempted to apply optical computing technology to radar systems, these studies still have certain limitations. For example, the Salvatore Maresca team (2020) introduced photonic systems into MIMO radars for the first time, expanding the coherence multiples of distributed radars. Experiments have shown that the input-multiple-output radar system implemented by photonics can maximize centralized data fusion and provide unprecedented resolution capabilities. However, the main contribution of this scheme is to improve the RF transmission link of distributed radars, using optical transmission RF to ensure coherence and resolution, but it does not involve the application of optical devices in signal processing, nor does it solve the high-throughput data processing problem brought about by centralized data fusion. Simin Li's team (2020) proposed a chip-based microwave photonic radar based on a silicon photonic platform. The integrated optical modulator completes the delinear frequency modulation of FMCW and obtains high-precision ranging results with a resolution of 2.7cm and an error of less than 2.75mm, and realizes inverse synthetic aperture imaging of multiple complex contour targets. However, this scheme is only applicable to FMCW radars, and multi-channel acquisition is still necessary, which fails to fundamentally solve the bottleneck problem of data processing. Although Bindong Gao's team (2021) used photogenerated RF technology to simultaneously generate M LFM transmission signals, and used photodecomposition linear frequency modulation technology to decompose N echo signals into MN intermediate frequency signals, thereby reducing the data sampling rate requirements, this method still requires processing a large amount of sampling channel data and fails to effectively reduce the overall data processing volume of the system.

[0004] In summary, although the application of optical computing technology in multi-channel forward-looking radar has achieved some initial results, the existing solutions still have many limitations, especially in the application of optical devices in signal processing and how to solve the problem of high-throughput data processing, which still needs further in-depth research and exploration. Summary of the invention

[0005] In view of the shortcomings of existing electrical architecture technology in processing high-throughput radar signals, the present invention proposes an optical processing device adapted to the parallel processing of multi-channel forward-looking radar signals, aiming to solve the processing bottleneck of high-throughput echo signals in multi-channel forward-looking radar systems. The device migrates the high-complexity links in the radar imaging algorithm to the optical domain for instantaneous parallel calculation, and compresses the amount of data during the processing process, thereby reducing the requirements for back-end electrical sampling and calculation, and realizing high-real-time radar forward-looking imaging. The device preprocesses high-throughput radar signals in the analog domain: the high-complexity links in the radar imaging algorithm are moved to the optical domain for instantaneous parallel calculation, and the amount of data is compressed during the processing process, and only the imaging time domain data required for subsequent electrical domain processing is retained, which reduces the requirements for back-end electrical sampling and calculation of imaging, and can realize high-real-time radar forward-looking imaging.

[0006] The technical solution of the present invention is as follows:

[0007] An optical processing device adapted for parallel processing of multi-channel forward-looking radar signals, used for parallel processing of N×M (M transmit, N receive) MIMO radar echo data, characterized in that it includes: M radar transmit array elements, N radar receive array elements, N down-conversion components, N light sources and modulators of different wavelengths, 2 N-channel wavelength division multiplexers, 2 optical delay arrays containing N independent branches, 1 N-channel wavelength division multiplexer, 2 one-dimensional optical convolution devices with levels of T0×f and Q, 2 optical switch devices, and 1 photoelectric detection device back-end electrical processor. Among them,

[0008] The receiving front end is composed of N radar receiving array elements and N down-conversion components, and the output N baseband signals are connected to the corresponding modulation units;

[0009] A light source array, consisting of N light sources with different wavelengths, connected to the local oscillator input end of the optical modulation array;

[0010] A multi-channel optical modulation array is composed of N modulators. The RF end of each modulator is connected to the baseband output port of the receiving front end with the corresponding number, and the local oscillator end is connected to the light source unit with the corresponding number. Light sources with different numbers have different wavelengths. It is used to modulate broadband echo signals into the optical domain for processing. The modulated output is connected to the optical delay array.

[0011] The first optical delay array is composed of N independent adjustable optical delay devices, the input of each delay device is connected to the output of the corresponding numbered optical modulator, the N delay devices all have a digital circuit control interface, the delay amount can be set through a program, and the delay output is connected to the wavelength division multiplexer;

[0012] A wavelength division multiplexer is used to combine N optical modulation signals of different wavelengths to achieve wavelength multiplexing, and the single-channel output after multiplexing is connected to the optical computing device array;

[0013] The first optical convolution kernel is composed of a T0×f tap FIR filter, which can perform convolution operation on the input N wavelength signals at the same time to realize pulse compression of radar signals, and the output is connected to optical switch 1;

[0014] The first optical switch has one input and one output, wherein the input is the output of the optical computing device, has a digital circuit control interface, and can control the on and off of the signal through a program, and its output is connected to the second optical computing device;

[0015] The second optical convolution kernel is composed of a Q-tap ​​FIR filter, which performs convolution calculations of fixed coefficients on the input N wavelength signals at the same time to realize pre-aliasing of analog signals, and the output is connected to the second optical switch;

[0016] The second optical switch has one input and one output, wherein the input is connected to the second output of the optical computing device, has a digital circuit control interface, and can control the on and off of the signal through a program, and its output is connected to the wavelength division multiplexing module;

[0017] The wavelength division multiplexing module separates a single channel of N optical modulated signals of different wavelengths into N independent branches. The output of the wavelength division multiplexer is connected to the optical delay array 2;

[0018] The second optical delay array is composed of N independent fixed optical delay devices, the input of each delay device is connected to the output of the wavelength division multiplexing module with a corresponding number, and the delayed output is connected to the optical combiner;

[0019] The optical combiner combines the delayed N optical signals of different wavelengths to form a time-division multiplexing frame signal corresponding to different receiving channels, and then outputs it to the photoelectric detection device;

[0020] The photoelectric detection device, which consists of a broadband photodetector, is used to convert the processed optical signal into an electrical signal, which is then transmitted to the back-end electrical processor.

[0021] The back-end electrical processor array consists of a high-speed ADC module and an electrical processor. The ADC samples and quantizes the optically processed signal, separates each frame, and stores it in different memory units through DMA. The processor first performs FFT on the frame signal to obtain the image spatial spectrum sampling value, and then fills the sampling value into an empty two-dimensional array according to the pre-calculated spatial spectrum position coordinates, and finally performs a two-dimensional IFFT on the array to obtain the radar forward imaging image.

[0022] Preferably, the wavelength interval of each light source in the light source array meets the multiplexing bandwidth requirement of the wavelength division multiplexer, and the wavelength uniqueness is achieved through digital tuning control.

[0023] Preferably, the first optical convolution kernel and the second optical convolution kernel are used to realize pulse compression of radar signals and pre-aliasing of analog signals, respectively, and the number of taps T0×f of the first optical convolution kernel is determined by the pulse width and bandwidth of the radar transmit signal, and is used to complete matched filtering in the optical domain; the number of taps Q of the second optical convolution kernel is equal to the data extraction multiple, and is used to complete time domain pre-aliasing in the optical domain, thereby reducing the amount of single-channel data by Q times.

[0024] Preferably, the first optical delay array and the second optical delay array are used to perform phase compensation and time domain delay on the optical modulated signal respectively, and the first optical delay array has an adjustable delay amount, and the second optical delay array has a fixed delay amount.

[0025] Preferably, the wavelength division multiplexer and wavelength demultiplexing module are used to realize wavelength multiplexing and separation of optical signals, and the wavelength division multiplexer combines N optical signals of different wavelengths into a single signal, and the wavelength demultiplexing module separates the single signal into N optical signals of different wavelengths.

[0026] Preferably, the time division multiplexing frame signal output by the optical combiner realizes non-overlapping combining of multi-channel signals through a fixed delay amount I×T1 / Q (I=0, 1, ..., N-1).

[0027] The technical advantages of the present invention are:

[0028] (1) It has ultra-wideband and high-throughput radar signal processing capabilities. The bandwidth is determined by the optical modulator and is not limited by the back-end ADC sampling rate, which is of great significance to the improvement of radar imaging resolution.

[0029] (2) It is able to complete the frequency domain extraction link in the imaging algorithm in the optical simulation domain, reducing the total back-end data volume of a single imaging to hundreds of times that of traditional solutions, while ensuring that the imaging accuracy is not inferior to that of traditional solutions.

[0030] (3) It can process multi-channel echo signals in parallel, giving full play to the architectural advantages of MIMO radar's "quick shot" imaging, and avoiding the disadvantages of traditional multi-channel acquisition and single-channel time-sharing acquisition solutions, such as high computational complexity, poor real-time performance, and high equipment cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A schematic diagram of an embodiment of an optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to the present invention, wherein N=8 and M=3;

[0032] Figure 2 The transformation of signal waveform during multi-channel data parallel processing and time domain extraction is demonstrated.

[0033] Figure 3 It is a schematic diagram of the process from sampling signal to image in the back-end electronic processor. DETAILED DESCRIPTION

[0034] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments, and a detailed implementation method and structure are given, but the protection scope of the present invention is not limited to the following embodiments. The "left" and "right" described below are based on the placement positions described in the figure, not the positions of the actual system. The actual system only needs to ensure that the connection order is the same as the figure.

[0035] An optical processing device adapted for parallel processing of multi-channel forward-looking radar signals is used to process multi-channel forward-looking radar signals composed of M transmitting units and N receiving units, with a transmitting signal length of T0, a signal acquisition window length of T1, a data rate reduction factor of Q, and a back-end ADC sampling rate of F. The device composition and functions are as follows:

[0036] The receiving front end 112 is composed of N independent receiving antennas and down-conversion components, and is used to receive radar echo signals and down-convert them into baseband signals; the baseband signal output by each receiving front end is connected to the corresponding modulation unit.

[0037] The light source array is composed of N light sources with different wavelengths and is used to generate a light sequence carrying a signal. Each light source is connected to the local oscillator input terminal of the corresponding modulator in the optical modulation array.

[0038] The optical modulation array consists of N modulators, and the RF end of each modulator is connected to the corresponding numbered baseband output port of the receiving front end. The modulator modulates the baseband signal into the optical domain for processing.

[0039] The first optical delay array is composed of N independent adjustable optical delay units and is used to perform phase compensation operation on the optical modulated signal. Each delay device has a digital circuit control interface, and the delay amount can be set by program.

[0040] The wavelength division multiplexer combines N optical modulation signals of different wavelengths to achieve wavelength multiplexing, and the multiplexed single-channel output is connected to the optical computing device array.

[0041] Optical computing device array (including a first optical convolution kernel, a first optical switch, a second optical convolution kernel, and a second optical switch):

[0042] The first optical convolution kernel is composed of a T0×f tap FIR filter, which is used to perform convolution operation on the input N wavelength signals to achieve pulse compression of radar signals;

[0043] A first optical switch, used for controlling the on and off of a signal, connected between the first optical convolution kernel and the second optical convolution kernel;

[0044] The second optical convolution kernel is composed of a Q-tap ​​FIR filter, which performs convolution calculations of fixed coefficients on the input N wavelength signals at the same time to achieve pre-aliasing of analog signals.

[0045] A second optical switch, used for controlling the on and off of the signal, connected between the second optical convolution core and the wavelength division multiplexing module;

[0046] The wavelength division multiplexing module separates a single channel of N optical modulation signals with different wavelengths into N independent branches.

[0047] The second optical delay array, composed of N independent fixed optical delay devices, is used to perform further delay operations on the separated optical signals.

[0048] The optical combiner combines N delayed optical signals of different wavelengths and time-division multiplexing to form time-division multiplexing frame signals corresponding to different receiving channels.

[0049] The photoelectric detection device, which consists of a broadband photoelectric detector, is used to convert the processed optical signal into an electrical signal and transmit it to the back-end electrical processor.

[0050] The back-end electrical processor array consists of a high-speed ADC module and an electrical processor. The ADC samples and quantizes the optically processed signal and stores each frame separately. The processor performs FFT and two-dimensional IFFT processing on the frame signal to obtain the radar forward imaging image.

[0051] Workflow:

[0052] The radar transmits a signal to illuminate the target, and the echoes of N receiving units (numbered I=0, 1, ..., N-1) are down-converted to baseband output. N independent light sources generate optical signals of different wavelengths, which are mixed with N baseband signals through N modulators, and the signals are modulated to the optical domain for processing. The N optical modulated signals are phase-compensated by N independent delay devices in the delay array. After phase compensation, the N optical modulated signals are combined into one multiplexed signal output through the N-channel wavelength division multiplexer. The multiplexed signal is pulse compressed by the first optical convolution kernel of length T0×f. After the convolution calculation, the first optical switch applies a rectangular window of time width T1 to the output signal. The signal after the range gate filtering is pre-aliased and extracted in the time domain by the convolution kernel 2 of length Q. After the convolution calculation, the second optical switch applies a rectangular window of time width T1 / Q to the output signal. After windowing, the optical signal is separated into N independent optical signals of different wavelengths through an N-channel wavelength division multiplexer, and the delay is set to I×T1 / Q according to the number of each signal. After delay, the N signals are combined through an N-channel wavelength division multiplexer, and finally converted to the electrical domain by a photodetector for final imaging.

[0053] The device embodiment realizes parallel processing of N-channel signals, reduces the single-channel data volume by Q times through extraction, reduces the multi-channel signal to a single-channel signal through delayed combination, reduces the number of ADC sampling channels to 1, and keeps the sampling rate unchanged. Assuming that the imaging resolution is set to n×n, after the optical domain front-end processing, the back-end electronic processor reduces the single-frame calculation to times.

[0054] See also Figure 1, when N=8, M=3, T0=3.2ns, T1=150ns, Q=15, F=20GHz, n=100, this embodiment includes 3 radar transmitting array elements, 8 radar receiving array elements, 8 down-conversion components, 8 light sources and modulators of different wavelengths, 2 8-channel wavelength division multiplexers, 2 optical delay arrays including 8 independent branches, 1 8-channel wavelength division multiplexer, 2 one-dimensional optical convolution devices with the order of T0×f and Q respectively, and 2 optical switch devices. The specific connection mode and function description of each device are as follows:

[0055] In the embodiment, the signal is outputted by three transmitting arrays 110 in time division to illuminate the forward-looking target. The echo is fed into the antenna feed system composed of eight receiving array elements 111, and is independently down-converted at the receiving front end 112 to obtain the corresponding baseband signal. The light source array 113 is used to generate an optical sequence carrying the signal, and the output wavelength of each light source 113 is different from each other, and in the embodiment, it is 8 different wavelengths. The optical power generated by the light source 113 is copied into 8 parts using a splitter at a ratio of 1:8, and is connected to the optical input ports of 8 modulators 114 respectively. The RF signal input port of the modulator receives the baseband signal obtained by down-conversion, and the output port of the modulator 114 is connected to the delay array 115. The delay array aligns the echo envelope, and the delay amount is the propagation delay of the negative receiving array element to the origin distance. The delay array output is connected to the input port of the wavelength division multiplexer array 116. The output port of the wavelength division multiplexer 116 is connected to the optical convolution kernel 117, where the number of levels of the convolution kernel is 64. The output of the optical convolution is connected to the optical switch 118, and a time domain window of T1=150ns is added to the output signal. The windowed output is connected to the optical convolution kernel 119, and the order of the convolution kernel is 15. The output of the convolution kernel is connected to the optical switch 120, and a time domain window of T2=10ns is added to the output signal. The output of the optical switch is connected to the 8-channel wavelength division multiplexer 121, and the optical signal is re-separated into 8 independent optical signals of different wavelengths through the 8-channel wavelength division multiplexer. The signal is connected to the delay array 122, and the delay amount is set to I×10ns (I=0,1,2,3,4,5,6,7) according to the number of each signal. After the delay, the signals do not overlap in the time domain, and the 8-way signals are combined through the 8-channel wavelength division multiplexer 123, and finally converted to the electrical domain through the photodetector 124. On the back-end electronic processor 125, the sampling is digitized through an ADC with a sampling rate of 20G, and each frame is separated according to the frame length of 10ns and stored in 8 different memory units through DMA. The processor first performs FFT on the 8 frames of signal to obtain the image spatial spectrum sampling value, and then fills the sampling value into an empty 100×100 two-dimensional array according to the pre-calculated spatial spectrum position coordinates. After all frames are filled, the array is subjected to a two-dimensional IFFT to obtain the radar forward imaging image.

[0056] The device embodiment realizes parallel processing of 8-channel signals, reduces the single-channel data volume by 15 times through extraction, reduces the multi-channel signal to a single-channel signal through delayed combination, reduces the number of electrical ADC sampling channels to 1, and the sampling rate is 20G. After optical domain front-end processing, the number of multiplications required for a single frame of the back-end electrical processor is reduced by 794,970 times, and the imaging time is reduced by 23 times.

[0057] Figure 2 The signal transformation process for realizing time domain extraction in an optical processing device is given when N=8, M=3, T0=3.2ns, T1=150ns, Q=15, F=20GHz, and n=100. Figure 3 The schematic diagram of the process from sampling signal to image in the back-end electronic processor is shown in the figure. As shown in the figure, the operations performed by the back-end electronic processor include: storing the signal sampled by ADC in frames; performing FFT on each frame of signal to generate spatial spectrum sampling values; filling the sampling values ​​into a preset two-dimensional array; and generating the radar forward imaging image through two-dimensional IFFT.

Claims

1. An optical processing device adapted for parallel processing of multi-channel forward-looking radar signals, used for parallel processing of N×M MIMO radar echo data, characterized in that: include: The receiving front end is composed of N radar receiving array elements and N down-conversion components, which are used to receive radar echo signals and down-convert them to baseband signals. The output N baseband signals are connected to the corresponding optical modulation arrays respectively. A light source array, consisting of N light sources with different wavelengths, each of which is connected to a local oscillator input terminal of the light modulation array; An optical modulation array, consisting of N modulators, wherein the RF end of each modulator is connected to the baseband output port of the corresponding receiving front end, and the local oscillator end is connected to the output end of the corresponding light source unit, for modulating the baseband signal into the optical domain; A first optical delay array, consisting of N adjustable optical delay units, wherein the input connection of each adjustable optical delay unit corresponds to the output of the modulator, is used to perform phase compensation on the optical modulated signal, and has a digital circuit control interface to set the delay amount; A wavelength division multiplexer is used to combine N optical modulation signals of different wavelengths into a single multiplexed signal; The first optical convolution kernel is composed of a FIR filter with T0×f taps, which performs pulse compression on the multiplexed optical signal; A first optical switch, used for controlling the on and off of a signal, connected between the first optical convolution kernel and the second optical convolution kernel; The second optical convolution kernel is composed of a FIR filter with Q taps, and performs pre-aliasing and extraction on the signal after being controlled by the first optical switch; The wavelength division multiplexing module separates the processed single-channel N optical modulation signals of different wavelengths into N independent branches; A second optical delay array, composed of N independent fixed optical delay units, is used to adjust the delay of the separated signal; The optical combiner combines the delayed N optical signals of different wavelengths into a time-division multiplexing frame signal; The photoelectric detection device is used to convert the combined optical signal into an electrical signal, transmit it to the back-end electrical processor array, and obtain the radar forward-looking image through the imaging algorithm.

2. The optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to claim 1, characterized in that: The wavelength interval of each light source in the light source array meets the multiplexing bandwidth requirement of the wavelength division multiplexer, and the wavelength uniqueness is achieved through digital tuning control.

3. The optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to claim 1, characterized in that: The first optical convolution kernel and the second optical convolution kernel are used to realize pulse compression of radar signals and pre-aliasing of analog signals, respectively. The number of taps T0×f of the first optical convolution kernel is determined by the pulse width and resolution requirements of the radar transmission signal, and is used to complete matched filtering in the optical domain. The number of taps Q of the second optical convolution kernel is equal to the data extraction multiple, and is used to complete time domain pre-aliasing in the optical domain, thereby reducing the single-channel data volume by Q times.

4. The optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to claim 1, characterized in that: The first optical delay array and the second optical delay array are used to perform phase compensation and time domain delay on the optical modulated signal respectively, and the first optical delay array has an adjustable delay amount, and the second optical delay array has a fixed delay amount.

5. The optical processing device for adapting multi-channel forward-looking radar signals according to claim 1, characterized in that: The wavelength division multiplexer and wavelength demultiplexing module are used to realize wavelength multiplexing and separation of optical signals. The wavelength division multiplexer combines N optical signals of different wavelengths into a single signal, and the wavelength demultiplexing module separates the single signal into N optical signals of different wavelengths.

6. The optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to claim 1, characterized in that: The time division multiplexing frame signal output by the optical combiner realizes non-overlapping combining of multi-channel signals through a fixed delay amount I×T1 / Q (I=0, 1, ..., N-1).

7. The optical processing device adapted for parallel processing of multi-channel forward-looking radar signals according to claim 1, characterized in that: The operations performed by the back-end electrical processor include: -Store the signal sampled by ADC in frames; -Perform FFT on each frame of signal to generate spatial spectrum sampling values; -Fill the sample values ​​into a preset two-dimensional array; -Generate radar forward-looking image through two-dimensional IFFT.

Images