High-precision correlation ranging system based on chaotic laser

Abstract

The invention discloses a high-precision correlation ranging system based on chaotic laser, and the system is characterized in that a transmitting end employs a VCSEL coherent array module, converts chaotic continuous light into pulse light and excites correlation mode locking through current pulse and orthogonal polarized light injection (OPIOI) modulation, and combines a photonic crystal coupler (hexagonal close arrangement, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection, orthogonal polarized light injection and orthogonal polarized light injection); the crystal lattice is 220nm / hole is 66nm), so that high-power and broadband TDS-free chaotic pulse synthesis is realized; a receiving end converts signals through a Geiger mode APD, an FPGA stores echoes in a BRAM in a segmented mode, cyclic shift XNOR operation and multi-segment cross-correlation accumulation are executed, time delay peak positioning is optimized in combination with a convolutional neural network, and finally a ranging value (the precision is superior to 1 micron) is output. The system has the advantages of anti-interference characteristic of chaotic laser and high resolution of pulse distance measurement, and supports USB serial port communication and real-time display of an upper computer. According to the invention, the problem of distance ambiguity of distance measurement by a phase method is solved, and meanwhile, the precision defect of a pulse method in dynamic target measurement is overcome.

Description

Technical Field

[0001] The present invention relates to related laser ranging technologies, belonging to the field of laser technology, and specifically relates to a high-precision correlation ranging system based on chaotic laser. It is applicable to medium and long-distance, high-precision dynamic target measurement scenarios, and the theoretical ranging accuracy is better than 1 micron. Background Art

[0002] Laser ranging technology is an active detection technology that can quickly and accurately obtain target distance information and has a very wide range of applications, including fields such as unmanned driving, satellite ranging, laser altimetry, underwater detection, and point cloud recognition. Traditional ranging methods include the pulse method, the phase method, etc., but traditional laser ranging technology has significant limitations.

[0003] The pulse method relies on high-peak-power pulses and has problems such as large system volume, low stability, and large measurement jitter for dynamic targets; the phase method is easily affected by environmental disturbances and produces a distance ambiguity effect; in order to measure longer distances and ensure high-precision information, the cross-correlation laser ranging method based on pseudo-random code modulation in the time domain has attracted much attention. By modulating the time-domain information of pulsed laser, after a long-distance delay, the main echo correlation peak information is matched to complete ranging. Under the existing pseudo-random code modulation cross-correlation ranging method, it is difficult to directly find the position of the code element with the largest correlation due to the deterioration of the signal-to-noise ratio in long-distance applications of the system; chaotic laser ranging realizes the design of a miniaturized laser ranging system with random time-domain signals, but is limited by time delay and insufficient bandwidth and is difficult to meet the high-precision ranging requirements. At the same time, the power of continuous chaotic laser is low and is blocked in long-distance measurements.

[0004] In order to find a ranging method that can achieve system miniaturization while taking into account the ranging range and accuracy in a high-noise environment. The present invention effectively overcomes the above defects through the chaotic mode locking of the VCSEL array and the TDS-free pulse synthesis of the photonic crystal coupler (output power ≥ 800 mW, bandwidth ≥ 20 GHz), combined with the multi-segment cross-correlation accumulation of the FPGA (cyclic shift exclusive NOR operation) and the CNN time delay correction (ΔD compensation). Summary of the Invention

[0005] The main differences between the present invention and the prior art are: integrating the VCSEL array, the photonic crystal coupler, and the FPGA module on the optical platform together, greatly reducing the volume of the laser ranging; the chaotic laser time-domain signal has natural non-periodicity and noise-like characteristics, belonging to a true random signal, without a fixed periodic signal sequence, reducing hardware operations such as random signal synchronization and period matching, saving resources, and having stronger anti-interference ability.

[0006] The photon counting ranging method that restores the time-domain signal to the digital level form uses the autocorrelation ranging scheme, but there is a theoretical ranging upper limit, and the theoretical ranging has no upper limit using the cross-correlation calculation method.

[0007] The method using SOA and optical fiber amplifier can achieve the amplification of chaotic power. However, due to the absence of pulse coherent mode locking, there is a maximum theoretical power amplification value, and the applicability is low. By using a photonic crystal coupler to achieve array coherent synthesis, high-power chaotic laser output can be realized.

[0008] The principle of realizing the technical difficulties of the present invention is as follows: The key point of the device is to first use a high-speed digital-to-analog converter (DAC) to generate a rectangular pulse sequence that satisfies the formula. The built-in timer module of the FPGA is used to generate a comb-like pulse sequence according to the repetition period Trep (10 ns). The high-speed digital convolution calculation is realized by using the DSP slice resources of the FPGA. The DAC is used to output a sine wave signal, and the modulation frequency f m In the range of 1 - 5 GHz, a polarization beam splitter and a λ / 2 wave plate are used to realize an injection optical path orthogonal to the main mode of the VCSEL array, and the injection power and frequency offset are dynamically adjusted to ensure that the injection light matches the frequency of the VCSEL cavity mode. Secondly, the photonic crystal selects an AlGaAs material with a relatively high refractive index to improve the light field confinement ability and reduce the coupling loss. The lattice period and aperture size error are controlled within 1% by electron beam lithography and reactive ion etching processes. Thirdly, the VCSEL array, photonic crystal, and amplification link are integrated on the same chip or packaging substrate to shorten the optical path and reduce power consumption. A thermoelectric cooling (TEC) module is used to precisely stabilize the temperature of key devices (such as VCSEL arrays, FPGAs) to compensate for environmental temperature changes. Rigid packaging and vibration damping brackets are used to meet the anti-vibration requirements in the industrial environment.

[0009] The key point of the method is the detection and correlation operation of the pulse signal of chaotic laser. Firstly, traditional chaotic laser is continuous light. A current pulse modulation is used to cut off the time-domain signal, which is convenient for analog-to-digital conversion and improves the accuracy and time of the correlation operation. Secondly, for the used correlation operation method, the key point is to perform a correlation operation between the echo time-domain signal recorded at the address and the main wave time-domain signal recorded initially. While reading and writing optical information data in the BRAM, the correlation operation is performed and the operation result is written into the address for the next round of BRAM reading and writing, which greatly improves the accuracy and operation rate compared with other traditional solution methods. Thirdly, the coherence calculation period processing is used to accumulate the digital sequences of 9 periods, increasing 2 9 For the actual signal statistical samples, it can be seen from the signal-to-noise ratio formula that the theoretical ranging accuracy is improved by 100 times, and at the same time, the data transmission time is theoretically only the hardware delay time, shortening the data processing time.

[0010] Aiming at the deficiencies existing in the existing laser technology, the purpose of the present invention is to provide a high-precision correlation ranging system based on chaotic laser, including the following modules:

[0011] (1) Transmission module: The 8×8 VCSEL array generates chaotic mode-locked pulses through current pulse and OPIOI dual modulation; coherent synthesis of the array is achieved through a photonic crystal coupler to generate a time-domain chaotic fluctuation envelope signal and output a broadband chaotic pulse without TDS.

[0012] (2) Reception and processing module: The PD and Geiger-mode APD respectively convert the main echo signal, and the FPGA stores the signal in segments in the dual-port BRAM; perform multi-segment cross-correlation operations (circular shift exclusive NOR + weighted average), and optimize the time-delay peak positioning through CNN.

[0013] (3) System integration: The coaxial optical path design (940nm laser + beam expander lens) reduces the volume of the beam splitter; USB serial communication (CP2103 chip) supports real-time display on the host computer (LabVIEW interface).

[0014] The specific technical solution includes the following steps:

[0015] Step 1: Adopt a vertical cavity surface emitting laser (VCSEL) array, and excite chaotic mode-locking through the synergistic effect of pulse current modulation and orthogonal polarization optical injection (OPIOI). The current modulation parameters are

[0016]

[0017] Among them, I th is the threshold current (300mA ≤ I th ≤ 500mA), ΔI is the chaotic perturbation amplitude (50mA ≤ ΔI ≤ 200mA), γ is the modulation depth (0.4 ≤ γ ≤ 0.8), f m is the modulation frequency (1GHz ≤ f m ≤ 5GHz), T w = 200ps (pulse width), T rep = 10ns (repetition period); the OPIOI injection power ≥ 5mW, the frequency offset is controlled within ±5GHz, and the wavelength deviation ≤ 0.02nm. Control the array element spacing to satisfy d ≤ λ / (2n eff ), where λ = 940nm, n eff = 3.2, ensuring that the optical field coupling efficiency ≥ 85%.

[0018] Step 2: Design a photonic crystal coupler using a hexagonal close-packed structure (lattice constant 220 nm, pore diameter 66 nm, AlGaAs material, refractive index 3.4) to achieve time-delay-free (TDS) coherent synthesis of the VCSEL array output. The coupling efficiency is ≥95%, the output power is ≥800 mW, and the spectral broadening is ≤0.05 nm. After being amplified by two stages of a semiconductor optical amplifier (SOA, gain 20 dB) and an erbium-doped fiber amplifier (EDFA, gain 15 dB), the chaotic pulse has a time jitter of ≤5 ps, providing a stable light source for high-precision ranging.

[0019] Step 3: Use an FPGA with an embedded dual-port BRAM unit (depth 16383, width 9) to complete the accumulation of multi-segment cross-correlation results and ping-pong read-write operations. The signal processing flow includes: storing the echo signal in segments according to the time width of the chaotic pulse in the BRAM; performing a cyclic shift exclusive NOR operation to generate a local correlation curve; weighted averaging the multi-segment results to suppress noise (suppression ratio ≥30 dB); locating the global peak of the cumulative curve to output the optimal time delay τ (resolution ≤0.3 ps). The pre-trained convolutional neural network (CNN) takes historical ranging data and the current peak position (τ, peak intensity, and adjacent point intensity) as inputs and outputs the environmental error correction amount ΔD, where the compensation error of ΔD is ≤0.1 μm.

[0020] As a further improvement of the present invention, the coaxial optical path of the system and the communication module adopt a coaxial integrated design of a transmitting end (940 nm laser + collimating / beam-expanding lens) and a receiving end (PD and Geiger-mode APD + signal shaping circuit), reducing the system volume by 50%. The FPGA realizes the UART-to-USB communication protocol (baud rate 9600 bps) through a CP2103 chip. The LabVIEW host computer software can display the ranging value and the dynamic target trajectory in real time and support the tracking of moving targets at a refresh rate of 1 kHz.

[0021] As a further improvement of the present invention, the ranging system module includes a 940 nm VCSEL array, an FPGA, a hardware conversion circuit, a photonic crystal coupler, an optical fiber amplifier, a collimating lens, and a beam-expanding lens;

[0022] The output end of the FPGA is connected to the hardware conversion circuit. Bank35 of the FPGA is configured with 8 pairs of LVDS differential outputs, which are directly connected to the VCSEL array driver board through an HDMI connector. The output end of the laser is combined with the collimating and beam-expanding lens to form a coaxial optical path.

[0023] As a further improvement of the present invention, the FPGA is selected as XC7A200T, and the hardware conversion circuit is an LVDS-to-pulse circuit and a radio frequency amplifier.

[0024] As a further improvement of the present invention, the laser received echo sequence is received by a Geiger APD, and the output end is connected to the signal shaping circuit, and the output end of the signal shaping circuit is connected to the input end of the FPGA.

[0025] As a further improvement of the present invention, the signal shaping circuit is a voltage amplifier and a voltage comparator.

[0026] As a further improvement of the present invention, the system further includes the following technical features: the dynamic threshold discrimination algorithm sets an adaptive threshold by calculating the average value of the photons stored in the BRAM, and the symbol restoration error rate ≤ 0.1%; the cross-correlation operation adopts a time isolation mechanism, so that the error between channels is only introduced by the symbol cross-correlation; the host computer fuses multi-period data to train the CNN model to optimize the ranging stability in complex environments.

[0027] As a further improvement of the present invention, the correlation detection calculation method is as follows: use the FPGA to perform a mathematical operation on the processed echo pulse sequence and the transmitted pulse sequence, shift each echo symbol one bit to the right in turn, and perform an exclusive NOR calculation on the corresponding symbols at each position of the shifted echo pulse sequence and the transmitted pulse sequence at the same time, and loop N times, compare the results of the N calculations, find the symbol position corresponding to the maximum value of the correlation calculation (theoretically 1), and upload it to the computer to measure the ranging distance.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] The laser ranging technology of the present invention adopts the self-correlation photon counting method. During the ranging process, three-level optoelectronic conversion is reduced by direct driving of the VCSEL, the ranging accuracy is improved to the sub-micron level (1μm level), the dynamic target tracking frequency reaches 1kHz, the anti-interference ability is enhanced (environmental light noise suppression > 30dB), and the system volume is reduced by 50%; a Geiger APD single-photon detector and a threshold discrimination algorithm are adopted, and a time isolation mechanism is adopted to make the error between different addresses only introduced by the symbol cross-correlation, significantly improving the signal-to-noise ratio of long-distance signals, with a ranging range ≥ 50km and an accuracy better than 0.15m; it supports medium and long-distance range measurement, has better comprehensive performance than the traditional pulse / phase method, takes into account both the ranging range and accuracy, and has a lower cost than the traditional scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a block diagram of the ranging system of the present invention.

[0031] Figure 2 It is a structural diagram of the VCSEL array and a schematic diagram of current pulse and optical injection modulation principle of the present invention.

[0032] Figure 3 It is a schematic diagram of the photonic crystal coupler and coherent synthesis of the present invention.

[0033] Figure 4 This is the logic block diagram of the FPGA for processing signal correlation operations of the present invention. Specific embodiments

[0034] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0035] The following will further describe the present invention in detail with reference to the accompanying drawings:

[0036] The present invention provides a high-precision correlation ranging system based on chaotic laser, which mainly performs optoelectronic double-time-domain modulation on chaotic laser signals, realizes coherent synthesis through a photonic crystal coupler to obtain high-power chaotic laser pulses, and realizes repeated transmission; during echo detection, signal trigger events within the pulse sequence time are recorded, the main echo discrete signals are read and written using BRAM, and a multiplier is used for correlation operations.

[0037] Specifically, the high-precision correlation ranging system and method based on chaotic laser in this embodiment include the following steps:

[0038] Step 1, the orthogonal polarization light injection (OPIOI) configures the seed light source wavelength to be 940 ± 0.02 nm (the deviation from the center wavelength of the VCSEL array ≤ 0.05 nm), the injection power ≥ 5 mW (accurately controlled by an attenuator), and the frequency deviation range -5 GHz ≤ Δf ≤ +5 GHz. This process is dynamically adjusted by a current tuning module; polarization control is performed, and a λ / 2 wave plate is used to adjust the polarization direction of the injected light to form a 90° angle with the main mode polarization of the VCSEL array, and real-time monitoring is performed through a polarization beam splitter to ensure that the polarization extinction ratio ≥ 20 dB.

[0039] Among them, orthogonal polarization light injection (OPIOI) is a key technology for realizing chaotic synchronous mode locking by injecting an optical signal with a specific wavelength and power in a manner orthogonal to the main mode polarization of the VCSEL array, combined with frequency deviation adjustment and polarization control.

[0040] Step 2, current pulse modulation is performed by adjusting the current,

[0041]

[0042] where I th is the threshold current (300 mA ≤ I th≤ 500 mA), ΔI is the chaotic perturbation amplitude (50 mA ≤ ΔI ≤ 200 mA), γ is the modulation depth (0.4 ≤ γ ≤ 0.8), f m is the modulation frequency (1 GHz ≤ f m ≤ 5 GHz), T w = 200 ps (pulse width), T rep = 10 ns (repetition period);

[0043] Among them, it should be noted that OPIOI is homologous with the current modulation clock (reference clock jitter ≤ 0.1 ps), and phase difference compensation is carried out through adjustable delay line calibration (calibration error ≤ 10 ps).

[0044] Step 3, start preparing the photonic crystal coupler. The substrate material is an AlGaAs wafer (refractive index 3.4, thickness 500 ± 10 nm). Electron beam lithography (EBL) is used to define the hole positions: lattice constant 220 ± 2 nm, hole diameter 66 ± 1 nm. After reactive ion etching (RIE), the perpendicularity deviation ≤ 1°, and the sidewall roughness ≤ 2 nm. The photonic crystal structure is hexagonal close-packed, as shown in the schematic Figure 3 figure.

[0045] Among them, it should be noted that it is necessary to detect by SEM to meet the hole array periodic error ≤ 1%, and use a near-field scanning optical microscope (NSOM) to verify the uniformity of the optical field distribution.

[0046] Step 4, perform optical path calibration to ensure that the parallelism between the output end face of the array unit and the input face of the coupler ≤ 0.01°. Use a piezoelectric ceramic displacement stage to compensate for the spot position offset; adjust the array unit spacing d ≤ λ / (2n eff ), and use an integrating sphere + power meter to monitor the synthesis efficiency of the device in real time, meeting the sampling rate of 1 kHz; achieve TDS-free coherent synthesis control.

[0047] Step 5, amplify the optical signal with InGaAs as the gain medium by SOA to achieve a gain of 20 dB and a saturated output power of 23 dBm. Temperature control is carried out through TEC refrigeration (stability ± 0.01 °C); use an EDFA with a pump power of 500 mW to achieve a chaotic pulse laser output with a power ≥ 800 mW and a spectral broadening ≤ 0.05 nm.

[0048] Step 6: The PD and Geiger APD collect chaotic signals and transmit them to the FPGA for waiting to be processed. An oscilloscope (bandwidth ≥ 20 GHz) can be used to observe the chaotic fluctuation envelope in the time-domain waveform. The on-chip dual-port BRAM of the FPGA is used as the operation unit for correlation accumulation. According to the chaotic time-domain signal, the single-cycle bandwidth is set to 10 ns, and the echo is read and written simultaneously. The depth of the dual-port BRAM is set to 16383 and the width is 9, which can achieve photon accumulation for up to 2^9 transmission cycles at most. The two ports, Port-A and Port-B, of the dual-port BRAM can control the data reading and writing of different address bits of the BRAM, which is very beneficial to the implementation of the ping-pong operation. After detecting the echo pulse, the frequency of the generated time-domain pulse envelope is greater than 1 GHz. Under the excitation of 100 MHz, the clock of the dual-port BRAM is divided into a cycle of 10 ns. Each clock needs to process one cycle of echo symbols. The steps to achieve correlation accumulation using the ping-pong operation are as follows: At the i-th clock, the correlation operation is performed on the echo time-domain signal recorded at address j and the main-wave time-domain signal recorded initially. The correlation peak value is written back to address j by Port-A to refresh the signal display. At the same time, the signal peak value stored at address j + 1 is read out by port-B. At the (i + 1)-th clock, after performing the correlation operation on the echo time-domain signal recorded at address j + 1 and the main-wave time-domain signal recorded initially, the obtained correlation time-delay peak value is written back to address j + 1 by Port-B to refresh the accumulated photon number. At the same time, the time-delay peak value recorded at address j + 2 is read out by Port-B... By operating in this cycle, at each clock, Port-A or Port-B processes one cycle of time-domain signal in address order to refresh the time-domain peak display. After refreshing the correlation peak value in address 16383, it returns to address 1 for operation again until all accumulation cycles are completed.

[0049] Furthermore, after completing the correlation data accumulation, in order to improve the robustness of the selected threshold, the FPGA calculates the mean value of the time-delay peak values stored in all address bits of the BRAM and sets it as the variable threshold to reduce the influence of the experimental environment on data processing. After the FPGA completes the selection of the variable threshold, it outputs the stored peak time delay and compares it with the threshold according to the address bits in turn, and uses the neural network algorithm for accuracy optimization.

[0050] Performing the correlation operation on the calculated echo period signal and the main-wave period signal, where the specific calculation method is:

[0051] The correlation operation is a process of obtaining distance information from the echo signal according to the basic characteristics of random signals. The correlation function can be expressed as:

[0052]

[0053] Discretize the random signal in the formula and perform Fourier operation. Through the multiplier, perform exclusive-NOR comparison on the signal sequence of each initial write address with the initial cycle pulse envelope signal stored in the FPGA to find the symbol position τ with the maximum correlation; the initial shift amount S of the echo signal is 0, and the total number of "1"s in the exclusive-NOR result with the main wave signal is the correlation intensity A0 at position 0. Let the initial value of the peak intensity be A m = A0, then circularly shift the echo signal to the higher bits to make S <= S + 1, calculate the correlation intensity A1 at position 1, and select the larger data between A1 and A m to refresh A m , repeat the above steps to obtain the final correlation intensity A m , the peak S m , the intensity A at position S m +1 m + , the intensity A at position S m -1 m - and other data.

[0054] Step 7, calculate the ranging distance of the target; the ranging distance D = τc / 2f, where f is the pulse interval repetition frequency.

[0055] Among them, the specific method is:

[0056] The upper computer completes the data stream transmission through the URAT to USB and USB to URAT serial port protocols, designs a basic software window through labview, adds a control ranging start button module, and a ranging information display module. By introducing a deep learning algorithm, the distance information is accumulated and optimized to the best value.

[0057] The CP2103 chip configured on the FPGA board can be used for the conversion of the UART to USB interface, enabling the upper computer on the PC side to establish communication with the FPGA, realizing the instruction issuance and data reception of the upper computer. After the upper computer establishes communication at a baud rate of 9600bps, it first sends a setting instruction to the FPGA. After waiting for the APD in the photon detector to be in the Geiger mode, the upper computer sends a start measurement instruction to the FPGA. After the FPGA receives the instruction, it starts ranging. Every time the FPGA completes a ranging, it transmits ranging data such as the relevant peak position, peak intensity, peak left and right point intensities, etc. to the upper computer through serial communication. The upper computer further accurately measures the ranging result according to the data using the convolutional neural network algorithm and completes the ranging display. After the upper computer sends a stop ranging instruction, the FPGA stops receiving the main wave signal and resolving the echo data; after the upper computer sends a system reset instruction, the FPGA stops ranging and restores the system to the state when it was just powered on.

[0058] Further, when the FPGA judges the content of the host computer instruction, it executes the corresponding instruction after it is consistent with the preset instruction code. When the FPGA judges the host computer instruction, first, two 8-bit registers reg1 and reg2 are used to store the high byte and the low byte transmitted by the serial port respectively. Whenever a new byte is received by the serial port, under the clock excitation, the data in reg1 is cyclically shifted to reg2, the data in reg2 overflows, and the new byte is stored in reg1. Finally, reg1 and reg2 are compared with the preset instruction code to obtain the instruction judgment result to be executed.

[0059] Further, after the lidar system completes a ranging, the FPGA performs data communication to send the ranging information back to the host computer. The sequence and format of the transmitted data are as follows:

[0060] (1) The lower 8 bits of the relevant peak position;

[0061] (2) The upper 8 bits of the relevant peak position;

[0062] (3) The lower 8 bits of the relevant peak intensity data;

[0063] (4) The upper 8 bits of the relevant peak intensity data;

[0064] (5) The lower 8 bits of the peak left point intensity data;

[0065] (6) The upper 8 bits of the peak left point intensity data;

[0066] (7) The lower 8 bits of the peak right point intensity data;

[0067] (8) The upper 8 bits of the peak right point intensity data;

[0068] (9) The lower 8 bits of the number of received code elements as the peak time delay;

[0069] (10) The middle 8 bits of the number of received code elements as the peak time delay;

[0070] (11) The upper 8 bits of the number of received code elements as the peak time delay;

[0071] The first 4 data each require 2 bytes, and the last data requires 3 bytes. A total of 11 bytes need to be sent for 5 data to transmit a ranging result to the host computer. After the host computer completes ranging, it counts the received data. When a total of 11 bytes from the FPGA are received, the bytes are combined in the order of first the low byte and then the high byte, byte 1 and byte 2 are combined, byte 3 and byte 4 are combined, byte 5 and byte 6 are combined, byte 7 and byte 8 are combined, and byte 9, byte 10, and byte 11 are combined to restore the data for further data processing.

[0072] Furthermore, the host computer fuses multi-cycle data, outputs the target motion trajectory, corrects the environmental error △D through a convolutional neural network, and achieves an accuracy better than 1μm.

[0073] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various changes and modifications can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A high-precision correlation ranging system based on chaotic laser, characterized in that, Including: System emission light source: A VCSEL coherent array module excited by pulse current and orthogonal polarization light injection OPIOI superposition modulation is used to achieve coherent mode locking of chaotic signals within the array. The chaotic signals show chaotic fluctuation envelopes in the time-domain waveform. Through a photonic crystal coupler, coherent synthesis of the VCSEL array is realized. After two-stage amplification by a semiconductor optical amplifier (SOA) and an erbium-doped fiber amplifier (EDFA), high-power chaotic laser pulses without time delay (TDS) and with a bandwidth ≥ 20 GHz are emitted. System receiving module: The main wave signal is received by a PD through a beam splitter, and the echo signal is collected by a Geiger-mode APD through a receiving lens and fiber coupling. The FPGA is used to segment and decompose the chaotic signal into interval random codes and store them in the BRAM unit. By performing cross-correlation operations on the read-write periodic random codes, the time-delay peak is located to obtain the time delay. After optimizing the results at different addresses through the convolutional neural network algorithm, the ranging value can be displayed, and the ranging accuracy is better than 1 μm.

2. The system according to claim 1, characterized in that: The method for the VCSEL coherent array to generate chaotic laser pulses includes: (a) Apply orthogonal polarization mode optical injection (OPIOI) to the VCSEL coherent array, with the injected optical power P inj ≥ 5 mW and the frequency offset Δν = ν inj - ν0 ∈ [-5, 5] GHz; the polarization direction of the orthogonal polarization mode optical injection (OPIOI) forms a 90° angle with the polarization direction of the main mode of the VCSEL array, and the wavelength deviation Δλ of the injected light ≤ 0.02 nm; inject seed light into the VCSEL array, with the wavelength deviation Δλ of the seed light from the center wavelength of the VCSEL array ≤ 0.05 nm and the polarization direction consistent with the main mode direction of the VCSEL array; (b) Inject a modulation current into each unit in the VCSEL array, and its expression is: Among them, I th is the threshold current, ΔI is the chaotic perturbation amplitude, γ is the modulation depth, f m is the modulation frequency, T w is the pulse width, T rep is the repetition period; t represents the time variable, with the unit of second (s); denotes the convolution of a rectangular pulse and a comb function, where comb is a comb function representing a sequence of pulses with an interval of Trep. (c) Control the array unit pitch d ≤ λ / (2n eff ), where λ represents the operating wavelength of the VCSEL in nanometers, and n eff represents the equivalent refractive index of the VCSEL material, which is a dimensionless parameter; this ensures that the optical field coupling efficiency is ≥ 85%.

3. The system according to claim 1, characterized in that: The photonic crystal coupler adopts a hexagonal close-packed structure with a lattice constant of 220 nm, a hole radius of 66 nm, and the material is AlGaAs. Initial chaotic pulses are generated by the VCSEL array, and after coherent synthesis by the photonic crystal coupler, the output power ≥ 800 mW. After two-stage amplification by the semiconductor optical amplifier (SOA) and the EDFA, high-power chaotic pulsed lasers are finally generated, with a spectral broadening ≤ 0.05 nm and a time jitter ≤ 5 ps.

4. The system according to claim 1, wherein: The FPGA has an internal dual-port BRAM unit with a depth of 16383 and a width of 9, and performs multi-segment cross-correlation result accumulation and ping-pong read-write operations. The time-delay peak is located by performing a cyclic shift exclusive NOR operation to obtain the time delay τ.

5. The system according to claim 1, wherein: The cross-correlation operation includes the following steps: (a) Segmentally store the chaotic signal into the BRAM, and the length of each segment matches the time width of the chaotic pulse. (b) Perform a cyclic shift exclusive NOR operation on each segment of the signal to generate a local correlation curve. (c) Accumulate the results of multiple segments and perform weighted averaging to suppress noise interference. (d) Locate the global peak of the accumulated curve and output the optimal time delay τ.

6. The system according to claim 5, characterized in that: The FPGA uses a pre-trained convolutional neural network, inputs historical ranging data and the current peak position, outputs the corrected distance ΔD, and finally the displayed distance is D = τc / 2 + ΔD, where c is the speed of light.

7. The system according to claim 1, characterized in that, The FPGA establishes a URAT-to-USB serial communication protocol with the PC through a CP2103 chip and uses the upper computer LabVIEW software for real-time display.

8. A ranging method based on the system according to any one of claims 1 to 7, characterized in that, Including the following steps: Step 1: Drive the VCSEL array to generate chaotic laser pulses, split the main wave signal, store one beam of the signal segmentally into the BRAM, and emit the other beam through a coaxial optical path to the target. Step 2: Receive the echo signal reflected by the target and store it segmentally into the BRAM, perform ping-pong read-write operations in the BRAM, and perform cross-correlation operations with the stored main wave signal. Step 3: Write the cross-correlation operation results into the refreshed address and accumulate the cross-correlation results of multiple segments. Step 4: Generate a symbol sequence through dynamic threshold discrimination and locate the delay peak τ; Step 5: Optimize and output the final distance D through the host computer neural network.

Images