Optical calculation real-time processing device for ultra-wideband radar imaging

Abstract

The invention relates to the technical field of ultra-wideband radar imaging and optical computing, in particular to an optical computing real-time processing device for ultra-wideband radar imaging, which comprises a signal input module, an optical convolution module, an optical adjustable delay line module and an electric processing module. On the basis of the advantage of ultrafast linear calculation of optical calculation, a core processing function in ultra-wideband radar imaging is effectively replaced, and the end-to-end processing time delay of the ultra-wideband radar imaging device is remarkably reduced. The end-to-end time delay bottleneck of traditional ultra-wideband radar imaging is broken through, a hardware implementation foundation is laid for an optical calculation real-time processing device oriented to ultra-wideband radar imaging, and an effective hardware solution is provided for ultra-wideband fine imaging application with high real-time performance requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of ultra-wideband radar imaging and optical computing technology, and particularly relates to an optical computing real-time processing device for ultra-wideband radar imaging. TECHNICAL BACKGROUND The ultra-wideband radar imaging technology plays a very important role in civil or military fields. Among them, the ultra-wideband synthetic aperture radar (SAR) is widely used in engineering exploration, battlefield detection and other scenes due to its high-resolution two-dimensional imaging capability of "all-weather, all-day, high penetration". The ultra-wideband synthetic aperture radar is called a high-resolution radar, mainly because of the high-resolution imaging in two dimensions of the range and azimuth. First, the synthetic aperture radar transmits an ultra-wideband signal, and through a specific matching operation at the receiving end, high-resolution imaging in the range direction is achieved, also known as fast-time dimension focusing. In the azimuth direction, to achieve slow-time dimension focusing, the relative motion between the radar and the target is relied on, and through multiple detections in the motion trajectory, echo signals from multiple azimuths are collected, and then multiple small real antenna apertures are combined into a large artificial equivalent antenna aperture with an effective antenna length of several kilometers by data processing. However, in order to avoid the spectral aliasing distortion of the ultra-wideband signal, the Nyquist sampling rate corresponding to the high instantaneous bandwidth in the fast-time dimension needs to be met; at the same time, in order to synthesize a large artificial equivalent antenna aperture to achieve high range resolution, a large amount of target reflection echoes need to be received in the slow-time dimension, which brings great digital domain calculation pressure and processing delay to the traditional digital imaging system.

[0002] However, due to the processing bottlenecks of the analog electronic system in terms of "working bandwidth, frequency range, system complexity", etc., it is difficult for the analog electronic system to directly perform finite impulse response filtering on the received ultra-wideband echo in the radio frequency domain, and only the radio frequency signal can be down-converted and then frequency domain matching filtering can be performed in the digital domain, which brings a large processing delay. At the same time, in order to avoid aliasing, a high sampling rate is generally required for the aliasing ultra-wideband signal, which makes it impossible to process a large amount of echo data in real time, and thus it is difficult to realize end-to-end low-latency ultra-wideband two-dimensional SAR imaging.

[0003] Accelerating the massive signal processing of ultra-wideband radar using real-time optical computing technology has become a highly promising approach to overcome the processing delay bottleneck of ultra-wideband radar imaging systems. In recent years, typical research results on optical real-time processing technology for ultra-wideband radar mainly involve the use of microwave photonics technology to construct ultra-wideband radar devices (Zhang Z, Liu Y, Stephens T, et al. Photonic radar for contactless vital sign detection[J]. Nature Photonics, 2023, 17(9): 791-797.) (Zhu S, Zhang Y, Feng J, et al. Integrated lithium niobate photonic millimetre-wave radar[J]. Nature Photonics, 2025, 19(2): 204-211.). The essence of its optical processing is to perform deskewing processing (frequency domain matched filtering) on ​​the echo signal. At the back end, the output signal of the optical processing system needs to be processed by time-frequency conversion, which still requires a certain processing delay and is difficult to meet the requirements of full real-time streaming processing. It is evident that existing optical computing processing devices for ultra-wideband radar still struggle to solve the end-to-end two-dimensional imaging latency problem, necessitating further research. Summary of the Invention

[0004] This invention addresses the performance limitations of existing ultra-wideband radar system processing devices by proposing a real-time optical computing processing device for ultra-wideband radar imaging. At the receiving end, range-direction pulse compression and azimuth-direction phase correction and focusing are performed on the echo signal. This invention selects the ultra-wideband LFM signal as the radar transmission signal and uses optical computing technology to physically replace the core functional modules in traditional devices, significantly overcoming the computational limitations of ultra-wideband SAR imaging devices and constructing a novel end-to-end time-domain processing device. The specific workflow of the device is as follows: A real-time optical computing processing device for ultra-wideband radar imaging, characterized in that it includes: The optical processing front-end, used for streaming processing of radio frequency echo signals in the optical domain, includes the following components connected in sequence: The signal input module is used to modulate the radio frequency echo signal into the optical domain; The optical convolution module is used to perform optical finite impulse response filtering on the electro-optically modulated echo signal to achieve range pulse compression. An optically adjustable delay line module is used to apply an adjustable delay to the pulse-compressed optical signal to compensate for phase errors in the slow time dimension. An electrical processing back end, connected to the optical processing front end, comprising: An optical-electrical conversion and digitization module for converting the processed optical signal into a digital electrical signal; A digital processor configured to perform a cyclic shift and two-dimensional projection coherent superposition algorithm on the digital electrical signal to output a radar image; Wherein the digital processor further provides a control signal to the optical adjustable delay module for dynamically adjusting the delay amount.

[0005] Further, the signal input module comprises a single-wavelength laser and a coherent modulator, which is a Mach-Zehnder modulator (MZM) working in a quadrature bias point mode to realize real signal optical domain bearing, or a double parallel Mach-Zehnder modulator (DPMZM) working in a single sideband suppressed carrier (SSB-SC) mode to realize complex IQ signal optical domain bearing.

[0006] Further, the complex IQ signal optical domain bearing scheme further comprises a cascaded frequency mixer before the DPMZM to realize integrated processing of optical modulation and IQ down-conversion.

[0007] Further, the optical convolution module comprises: A cascaded N optical splitters for splitting each input optical signal into two output signals; N optical branches in parallel, each branch comprising an optical delay line and a weight regulator; A cascaded N beam combiners for coherently combining the output signals of the N branches; Wherein the delay amounts of the optical delay lines are equally spaced, and the weights of the weight regulators are set according to the coefficients of the pulse compression matched filter.

[0008] Further, the splitting ratio of the optical splitter is set to 1:(N-1) to 1:1, and the delay amounts are equally spaced according to the sampling interval Ts; the weights of the weight regulators are set to the discrete sampling values of the matched filter function, and the splitting ratios of the N beam combiners are 1:1, 1:2, …, 1:(N-1), 1:N respectively. 1)。

[0009] Further, the optical splitter and the beam combiner can be configured in a form of adjustable splitting ratio, without the weight regulator, and the splitting and weighting operations are completed simultaneously by adjusting the splitting coefficients.

[0010] Further, the weights of the weight regulator can be statically adjusted by a voltage signal in advance, so that the weight coefficients of the branch optical power satisfy the discrete value configuration of the matched filter, to realize the weight configuration of the pulse compression convolution.

[0011] Further, the optical adjustable delay line module is a programmable optical fiber delay line based on a semiconductor optical amplifier switch array, or an integrated optical delay chip based on waveguide technology, or an electrically adjustable optical delay line.

[0012] Further, the electrical processor feeds back a voltage value to the optical adjustable delay line module (300) through an electrical output port, which can be provided by a backend specific parameter estimation algorithm.

[0013] Further, the sampling mode of the electrical processor includes radio frequency direct sampling or sampling of a baseband signal after frequency down conversion.

[0014] Further, the system can be composed of discrete devices, or an integrated chip platform, or a system using both discrete components and integrated chips. When using discrete devices, the connections between modules and between devices within a module can be achieved through optical fibers, spatial optical devices, or other methods that limit the propagation path of light beams. If an integrated chip platform is used, the connections between components can be achieved using integrated waveguides.

[0015] The signal modulator, beam splitter, weight controller, and combiner used in the signal input module and optical convolution module are usually based on a coherent light Mach-Zehnder interferometer structure, which uses the Pockels effect of lithium niobate and the carrier dispersion effect to achieve optical phase shift and modulation by changing the optical refractive index through external energy. Further, the weight controller can also use optical devices such as micro-ring resonators, electro-absorption modulators, thermo-optic modulators, and electro-optic modulators that have amplitude adjustment effects on light fields to achieve the same weighted multiplication effect.

[0016] The optical delay line can achieve delay effect through the propagation time of light. The methods for achieving optical signal delay include optical fiber delay line, optical waveguide delay line, slow light delay line, or spatial free propagation delay, etc. Light access and other components that use storage time to achieve delay effect can also be used.

[0017] The technical advantages of the present application are: The present application uses optical convolution devices and optical adjustable delay line devices to directly complete the stream processing of pulse compression and phase compensation in the radio frequency domain for received ultra-wideband radar signals, and with a small amount of electrical backend processing, end-to-end real-time ultra-wideband two-dimensional imaging can be achieved. This provides a unique and effective solution to break through the time delay of end-to-end two-dimensional imaging of ultra-wideband radar devices, and is widely applicable to the field of ultra-wideband radar imaging applications for civilian or military use. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1The figure is an embodiment of the optical computing real-time processing device for ultra-wideband radar imaging of the present application. The connection mode of a feasible optoelectronic processing module for realizing the ultra-wideband radar imaging of the present application and the hierarchy of each module are shown.

[0019] Figure 2 The figure is an implementation scheme of the optical domain bearing device for radar signals. It comprises Figure 2 (a) The figure shows a direct real signal modulation scheme, which uses a single MZI modulator to load a baseband signal to the amplitude of an optical carrier; Figure 2 (b) The figure shows a complex signal optical domain processing scheme, which realizes optical modulation and IQ demodulation through a DPMZM or a joint mixer working in SSB-SC mode.

[0020] Figure 3 The figure is an optical two-dimensional convolution device, which is Figure 1 an alternative implementation device of the optical convolution device.

[0021] Figure 4 The figure is a preferred hardware implementation device of the optical adjustable delay line device. DETAILED DESCRIPTION

[0022] The technical solutions of the present application will be described in detail below in combination with the drawings and embodiments, and a detailed implementation manner and structure are given, but the protection scope of the present application is not limited to the following embodiments.

[0023] Please refer to Figure 1 , Figure 1 The figure is an embodiment of the optical computing real-time processing device for ultra-wideband radar imaging of the present application, which comprises a signal input module 100, an optical convolution module 200, an optical adjustable delay line module 300 and an electrical processing module 400. The processing flow and the functions of each module are tightly coupled, and the details are as follows: The signal input module 100 has one electrical input port and one optical output port. The analog domain radio frequency signal received by the antenna is input through the electrical input port, and the optical output port is connected to the input port of the subsequent optical convolution module. The signal input module is composed of a receiving antenna 101, a single-wavelength laser 102 and a coherent modulator 103. The single-wavelength laser 102 serves as a light source, and the coherent modulator 103 works in the quadrature bias point mode, which is responsible for linearly modulating the radio frequency signal input by the receiving antenna 101 to the optical amplitude. After passing through the module, the value of the received microwave signal will be represented by the optical amplitude value, and the signal will be transmitted to the subsequent optical convolution module 200.

[0024] The optical convolution module 200 has one optical input port, one optical output port and N electrical input ports. The optical input port of the optical convolution module 200 is connected in series with N optical splitters 201, and the splitting ratios are 1:(N 1), 1:(N 2),…,1:1, each beam splitter 201 outputs two optical paths, which are connected to the optical delay line 202 and the weighting regulator 203 respectively. The optical delay lines 202 are also arranged in series, and the delay amount of each delay line is consistent, corresponding to the discrete time delay interval T of the matched filter tap. s Through this arrangement, the output optical signal of the other connection weight controller 203 of the cascaded optical splitter can be made to have 0, T respectively. s ,…,(N-1)* T s Different delay amounts. Each weight regulator 203 is connected to one electrical signal input port. Due to the splitting ratio parameter setting of the beam splitter 201, each branch connected to the weight regulator 203 has the same optical signal power. The weight value (discrete value of the matched filter) corresponding to the weight regulator 203 can be changed through the electrical input port to adjust the optical signal power of the corresponding branch, which is equivalent to completing the weight multiplication calculation. Furthermore, the output port of each weight regulator 203 is connected to the input port of the cascaded beam combiner 204. The splitting ratio of the beam combiner 204 is 1:1, 1:2, ..., 1:(N) respectively. 1) The accumulation of optical signals is completed, thereby completing the complete calculation process corresponding to optical pulse compression.

[0025] The optically adjustable delay line module 300 has one optical input port, one optical output port, and one electrical input port. The optical input port of the optically adjustable delay line module 300 is connected to the optical convolution module 200, and the electrical input port and optical output port are connected to the electrical processing module 400. The electrical input port receives feedback parameters from the back-end electrical processing module 400 to adjust the optical delay. The optical output port transmits the signal output from the optical processing device to the back-end electrical processing module 400 to complete the remaining electrical processing operations.

[0026] In this embodiment, the electrical processing module 400 has one optical input port and one electrical output port. The optical input port is connected to a photodetector 401 to convert photoelectric signals and output them to the electrical processor 402. The electrical output port is connected to the electrical input port of the optical adjustable delay line module 300. The electrical processor 402 is responsible for performing minor processing tasks such as cyclic shifting and two-dimensional coordinate transformation during imaging, and for achieving final coherent superposition and image display.

[0027] Please see Figure 2 , Figure 2 The hardware implementation methods for optical transmission of radar signals include (a) direct transmission scheme and (b) IQ complex transmission scheme.

[0028] against Figure 1The signal input module 100 can be configured as follows: Figure 2 As shown in (a), the real signal input from the receiving antenna is amplified by a low-noise amplifier and then directly input to the MZM. Another method is as follows... Figure 2 As shown in (b), the signal is input through the receiving antenna, amplified by a low-noise amplifier, and then split into two paths by a power divider. One path is applied to the upper arm of the DPMZM, and the other path is applied to the lower arm of the DPMZM after passing through a 90° phase-shifting bridge, thus achieving IQ complex load. Furthermore, an orthogonal mixer can be added after the power divider to demodulate the two signals separately, achieving IQ down-conversion.

[0029] Please see Figure 3 , Figure 3 Another hardware implementation of an optical pulse compression device includes a device arrangement structure for an optical two-dimensional convolution device.

[0030] Figure 1 The optical convolution module in the middle can be replaced by Figure 3 The two-dimensional optical FIR form shown. Figure 3 The device shown can distribute signals with different delays through cascaded beam splitters and first-stage delay lines. Within each branch, a beam splitter further divides the signal into N sub-paths, each connected to a weighting regulator. Subsequently, the sub-paths are cross-combined between branches, and the combined results are accumulated using a beam combiner. Finally, a second-stage delay line and a cascaded beam combiner are used to accumulate the delays between different combinations, where the delay ratio of the first-stage delay line to the second-stage delay line is... : or : ,in The convolution scale, M, represents the number of branches in the two-dimensional convolution. This two-dimensional FIR structure significantly reduces the number of delay lines required to achieve the same convolution scale, thereby reducing overall architecture loss and significantly improving the scalability of optical convolution scale.

[0031] Please see Figure 4 , Figure 4 for Figure 1 One hardware implementation method for an optical delay line module includes a hardware architecture for an N-level programmable optical delay line.

[0032] Specifically, targeting Figure 1 The hardware implementation architecture of the optically adjustable delay line module 300 can be as follows: Figure 4As shown. Including a programmable optical switch controlled by a cascade delay line, containing N stages, each stage delay line can be controlled by optical switch whether to access the cascade delay link, so as to realize the real-time selection of Nbit delay amount, corresponding to the implementable radar signal phase adjustment. The delay precision and stage number of the optical delay line can be adjusted to meet different phase precision requirements. In addition, in the scene with low real-time requirement, the equivalent function can also be realized by using electrically driven optical delay line instead.

[0033] In summary, the application uses optical computing technology to realize the equivalent replacement of the core processing function in the ultra-wideband SAR imaging device, and through the effective cooperation of the optoelectronic device, develops an optical computing real-time processing device for ultra-wideband radar imaging, which can have extremely low end-to-end delay under the condition of ensuring normal imaging quality, and has high practical value in some imaging scenes with high requirements for resolution and time delay.

Claims

1. An optical computing real-time processing device for ultra-wideband radar imaging, characterized in that, The application relates to a radar signal processing system, which comprises: an optical processing front end for streaming radio frequency echo signals in the optical domain, which comprises, in sequence: a signal input module (100) for modulating radio frequency echo signals into the optical domain; an optical convolution module (200) for performing optical finite impulse response filtering on the echo signals after electro-optical modulation to realize range direction pulse compression; an optical adjustable delay line module (300) for applying an adjustable delay amount to the pulse compressed optical signals to compensate for phase errors in the slow time dimension; an electrical processing back end connected to the optical processing front end, which comprises: an optical-electric conversion and digitization module for converting the processed optical signals into digital electrical signals; a digital processor configured to perform a circular shift and two-dimensional projection coherent superposition algorithm on the digital electrical signals to output a radar image; wherein the digital processor further provides a control signal to the optical adjustable delay module for dynamically adjusting the delay amount.

2. The photonic computing real-time processing device of claim 1, wherein, The signal input module (100) comprises a single-wavelength laser (101) and a coherent modulator (102), wherein the coherent modulator (102) is a Mach-Zehnder modulator (MZM) working in a quadrature bias point mode to realize real signal optical domain bearing, or a double parallel Mach-Zehnder modulator (DPMZM) working in a single sideband suppressed carrier (SSB-SC) mode to realize complex IQ signal optical domain bearing.

3. The photonic computing real-time processing device of claim 2, wherein, The complex IQ signal optical domain bearing scheme further comprises a cascaded frequency mixer before the DPMZM to realize integrated processing of optical modulation and IQ down-conversion.

4. The photonic computing real-time processing device of claim 1, wherein, The optical convolution module (200) comprises: N cascaded optical splitters (201) for splitting each input optical signal into two output optical signals; N parallel optical branches, each of which comprises an optical delay line (202) and a weight regulator (203); N cascaded beam combiners (204) for coherently combining the output signals of the N branches; wherein the delay amounts of the optical delay lines (202) are equally spaced, and the weights of the weight regulators (203) are set according to the coefficients of the pulse compression matched filter.

5. The photonic computing real-time processing device of claim 4, wherein, The splitting ratio of the optical splitter (201) is set to 1:(N-1) to 1:1, and the delay amount is configured with equal intervals of sampling interval Ts; the weight value of the weight value regulator (203) is set to the discrete sampling value of the matched filter function, and the splitting ratios of the N beam combiners (204) are 1:1, 1:2, …, 1:(N-1) in turn 1).

6. The photonic computing real-time processing device of claim 4, wherein, The optical splitters (201) and the beam combiners (204) can be configured in the form of adjustable splitting ratios, and the weight regulators (203) are not needed, and the splitting and weighting operations can be simultaneously completed by adjusting the splitting coefficients.

7. The photonic computing real-time processing device of claim 4, wherein, The weights of the weight regulators (203) can be statically adjusted by voltage signals in advance, so that the weight coefficients of the branch optical powers meet the discrete numerical configuration of the matched filter, so as to realize the weight configuration of the pulse compression convolution.

8. The photonic computing real-time processing device of claim 1, wherein, The optical adjustable delay line module (300) is a programmable optical fiber delay line based on a semiconductor optical amplifier switch array, or an integrated optical delay chip based on waveguide technology.

9. The photonic computing real-time processing device of claim 1, wherein, The electrical processor (401) feeds back a voltage value to the optical adjustable delay line module (300) through an electrical output port, and the voltage value can be provided by a back-end specific parameter estimation algorithm.

10. The photonic computing real-time processing device of claim 1, wherein, The optical convolution module (200) is generally based on a coherent optical Mach-Zehnder interferometer structure, and also includes other analog domain optical convolution calculation architectures composed of micro-ring resonators, electro-absorption modulators, thermo-optic modulators, electro-optic modulators, and the like.

11. The photonic computing real-time processing device of claim 1, wherein, The connection sequence of the optical convolution module (200) and the optical adjustable delay line module (300) can be adjusted.

12. The photonic computing real-time processing device of claim 1, wherein, The system can be composed of discrete devices, or an integrated chip platform, or a system using discrete components and integrated chips in combination. When using discrete devices, the connections between modules and between devices within a module can be achieved through optical fibers, spatial optical devices, or other methods of limiting the propagation path of light beams. If an integrated chip platform is used, the connections between components can be achieved using integrated waveguides.

Images