Autonomous vehicle, autonomous vehicle control system, and light detection and ranging system

Abstract

An autonomous vehicle, an autonomous vehicle control system, and an optical detection and ranging system are disclosed. The autonomous vehicle control system includes: one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to: receive an electrical signal generated based on a reflected optical signal from an object; transform the electrical signal into the frequency domain to generate a transformed electrical signal; determine the Doppler frequency shift of the reflected optical signal based on the cross spectrum of the transformed electrical signal; determine movement information based on the Doppler frequency shift indicating whether the object is moving closer to or further away from the autonomous vehicle control system; and control at least one of a steering system or a braking system based on the movement information.

Description

[0001] This application is a divisional application of National Application No. 202080051843.8 (International Application No. PCT / US2020 / 041809, International Application Date July 13, 2020, Invention Title "Method and System for Enhanced Velocity Resolution and Signal-to-Noise Ratio in Distance Detection Using Optical Phase Coding")

[0002] Cross-reference to related applications

[0003] This application claims the benefit and priority of U.S. Provisional Patent Application No. 62 / 874835, filed July 16, 2019, the entire disclosure of which is incorporated herein by reference. Technical Field

[0004] Various aspects of this disclosure generally relate to optical detection and ranging (LIDAR) in the field of optics, and more specifically to systems and methods for enhancing velocity resolution and signal-to-noise ratio in distance detection using optical phase coding. Background Technology

[0005] Optical distance detection, using lasers (often referred to by the mnemonic LIDAR), is used in a variety of applications, from altitude measurement to imaging to collision avoidance. LIDAR provides finer-scale distance resolution with a smaller beam size than conventional microwave ranging systems such as Radio Detection and Ranging (RADAR). Optical distance detection can be achieved using several different techniques, including direct ranging based on the round-trip time of the optical pulse to the object, linear frequency modulation (LFM) detection based on the frequency difference between the transmitted chirped optical signal and the reflected signal scattered from the object, and phase-coded detection based on a sequence of single-frequency phase changes distinguishable from natural signals. Summary of the Invention

[0006] One embodiment disclosed herein relates to a system for enhanced velocity resolution and signal-to-noise ratio in distance detection using optical phase coding. In some embodiments, the system includes receiving an electrical signal generated by mixing a first optical signal and a second optical signal, wherein the first optical signal is generated by modulating an optical signal, and wherein the second optical signal is received in response to sending the first optical signal toward an object. In some embodiments, the system includes determining a Doppler shift of the second optical signal. In some embodiments, the system includes generating a corrected electrical signal by adjusting the electrical signal based on the Doppler shift. In some embodiments, the system includes determining the distance to the object based on a cross-correlation associated with the corrected electrical signal.

[0007] In another aspect, this disclosure relates to an optical detection and ranging (LIDAR) system for enhanced velocity resolution and signal-to-noise ratio in optical phase-coded distance detection. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to receive an electrical signal generated by mixing a first optical signal and a second optical signal, wherein the first optical signal is generated by modulating an optical signal, and wherein the second optical signal is received in response to sending the first optical signal toward an object. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to determine a spectrum over a first duration of the electrical signal. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to determine a Doppler shift of the second optical signal based on the spectrum. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to generate a corrected electrical signal by adjusting the electrical signal based on a Doppler frequency shift. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to determine the distance to an object based on a cross-correlation of a first duration of the corrected electrical signal and a second duration of a phase-coded radio frequency (RF) signal associated with a first optical signal, the first duration being different from the second duration.

[0008] In another aspect, this disclosure relates to an autonomous vehicle including a Light Detection and Ranging (LIDAR) system. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to receive an electrical signal generated by mixing a first optical signal and a second optical signal, wherein the first optical signal is generated by modulating the optical signal, and wherein the second optical signal is received in response to sending the first optical signal toward an object. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to determine a Doppler frequency shift of the second optical signal. In some embodiments, the LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to generate a corrected electrical signal by adjusting the electrical signal based on the Doppler frequency shift. In some embodiments, a LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to determine the distance to an object based on a cross-correlation between a corrected electrical signal and a radio frequency (RF) signal associated with a first optical signal. In some embodiments, a LIDAR system includes one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to operate an autonomous vehicle based on the distance to an object.

[0009] Other aspects, features, and advantages will become apparent from the following detailed description, simply by illustrating numerous specific embodiments, including the best mode contemplated for implementing the embodiments of this disclosure. Other embodiments are also capable of having other and different features and advantages, and several details thereof may be modified in various obvious ways, all without departing from the spirit and scope of the embodiments. Therefore, the drawings and descriptions should be considered illustrative in nature, rather than limiting. Attached Figure Description

[0010] The embodiments are illustrated by way of example and not limitation in the accompanying drawings, in which similar reference numerals denote similar elements, and wherein:

[0011] Figure 1A This is a schematic diagram illustrating an example of transmitting an optical phase-coded signal for distance measurement according to an embodiment;

[0012] Figure 1BThis illustrates a series of binary numbers for distance measurement according to an embodiment. Figure 1A A schematic diagram illustrating the transmission and return of optical signals;

[0013] Figure 1C This is a schematic diagram illustrating an example cross-correlation between a reference signal and two return signals according to an embodiment;

[0014] Figure 1D This is a schematic diagram showing an example spectrum of a reference signal and an example spectrum of a Doppler shift return signal according to an embodiment;

[0015] Figure 1E This is a schematic diagram illustrating an example cross spectrum of the phase component of the Doppler frequency shift return signal according to an embodiment;

[0016] Figure 2 This is a block diagram illustrating example components of a high-resolution LIDAR system according to an embodiment;

[0017] Figure 3A This is a block diagram illustrating example components of a phase-coded LiDAR system according to an embodiment;

[0018] Figure 3B This is a block diagram illustrating example components of a Doppler-compensated phase-coded LIDAR system according to an embodiment;

[0019] Figure 4A This is a flowchart illustrating an example method for determining and compensating for the Doppler effect on distance using a Doppler-corrected phase-coded LIDAR system according to an embodiment;

[0020] Figure 4B This is a flowchart illustrating an example method for enhancing velocity resolution and signal-to-noise ratio in optical phase-coded distance detection according to an embodiment;

[0021] Figure 5A This is a block diagram illustrating an example of multiple time blocks according to an embodiment and a first duration longer than each time block, wherein each time block is Figure 1B The duration of the phase code;

[0022] Figure 5B It is a graph showing the relationship between an example power spectrum according to an implementation method and frequencies calculated for each time block and for a first duration in multiple time blocks;

[0023] Figure 5C This illustrates an embodiment. Figure 5A A block diagram illustrating multiple time blocks and multiple time periods of a first duration, where consecutive time periods overlap;

[0024] Figure 6A This is a distance distribution diagram showing the actual distance peaks, which are almost indistinguishable in terms of power from noise ranges at greater distances, according to an embodiment.

[0025] Figure 6B It is a graph of the FFT independent variable in Equation 16b, based on the real and imaginary parts of the return signal according to the implementation method.

[0026] Figure 7 This is a graph illustrating an example dependence of the phase compensation complex value of the distance peak with respect to the symbol of the Doppler frequency shift according to an embodiment;

[0027] Figure 8 This is a block diagram illustrating an example system 801 including at least one high-resolution Doppler LIDAR system 820 mounted on a vehicle 810 according to an embodiment;

[0028] Figure 9A and Figure 9B This is a graph showing the reduction in SNR caused by laser linewidth according to an embodiment, without compensation, for two different coherent processing intervals of 2 μs and 3 μs distance and sampling rate.

[0029] Figure 10A This is a graph showing example distributions of signal and noise in analog data with various linewidth corrections according to various implementations;

[0030] Figure 10B and Figure 10C This is a graph showing example distance peaks in actual returned data from various implementations that apply different linewidth corrections;

[0031] Figure 11A and Figure 11B The spectrum illustrates the example effects of frequency broadening due to speckle on the selection of the Doppler peak at two different sampling rates, according to various embodiments.

[0032] Figure 11C and Figure 11D This is a distance diagram illustrating example effects of digital compensation for frequency broadening for two different sampling rates according to various embodiments;

[0033] Figure 11E and Figure 11F This is a graph illustrating an example improvement in signal-to-noise ratio as a function of distance, resulting from digital compensation for frequency broadening with two different choices of the number of Doppler peaks used, according to various embodiments.

[0034] Figure 12 This is a block diagram illustrating a computer system on which embodiments of the present disclosure may be implemented; and

[0035] Figure 13 A chipset on which embodiments of the present disclosure can be implemented is shown. Detailed Implementation

[0036] To achieve acceptable range accuracy and detection sensitivity, direct long-range LIDAR systems use short-pulse lasers with low pulse repetition rates and extremely high peak power. High pulse power can lead to rapid degradation of the optical components. Linear frequency modulated (LFM) and phase-coded LIDAR systems use long optical pulses with relatively low peak power. In this configuration, range accuracy increases with the length and bandwidth of the LFM bandwidth or phase coding, rather than the pulse duration, and therefore excellent range accuracy can still be achieved.

[0037] Useful optical linear frequency modulation (LFM) bandwidths have been achieved by modulating optical carriers with broadband radio frequency (RF) electrical signals. Recent advances in LFM LiDAR involve using the same modulated optical carrier as a reference signal, which is combined with the return signal at an optical detector to generate a relatively low beat frequency in the resulting electrical signal within the RF band. This beat frequency is proportional to the frequency or phase difference between the reference and return optical signals. This beat frequency detection of the frequency difference at the detector is called heterodyne detection. It has several advantages known in the art, such as the use of readily available and inexpensive RF components. The recent work described in Patent 7,742,152 illustrates a novel and simpler arrangement of optical components using an optical signal separated from the transmitted optical signal as a reference optical signal. Except for terminology inconsistent with that used herein, the entire contents of this patent are incorporated herein by reference as if fully set forth herein. This arrangement is referred to in this patent as heterodyne detection.

[0038] LiDAR detection with phase-coded microwave signals modulated onto an optical carrier has also been used. Here, the bandwidth B is proportional to the reciprocal of the duration τ of the pulse used to carry each phase (B = 1 / τ), where any phase-coded signal consists of a large number of such pulses. This technique relies on correlating a sequence of phases (or phase changes) at specific frequencies in the returned signal with a sequence in the transmitted signal. The time delay associated with the peaks in the correlation is related to the distance via the speed of light in the medium. The distance resolution is proportional to the pulse width τ. The advantages of this technique include the need for fewer components and the use of mass-produced hardware components developed for phase-coded microwave and optical communications.

[0039] However, phase-coded LIDAR systems that implement the aforementioned methods for scribed Doppler detection often struggle to provide target velocity resolution suitable for autonomous vehicle (AV) applications.

[0040] Therefore, this disclosure relates to systems and methods for enhancing the performance of LiDAR. Specifically, this disclosure describes a LiDAR system with a synchronous processing arrangement in which a transmitted optical signal and a reference optical signal are generated from the same carrier, resulting in a phase correlation between the Doppler frequency shift signal and the range signal. This provides significant advantages in compensating for multiple Doppler signals from a single target, eliminating range signals that are not considered due to inconsistent phase, and determining the sign of the Doppler velocity from the real-valued signal as the sign shift of the correlated phase.

[0041] Furthermore, this invention describes a LiDAR system with an asynchronous processing arrangement. That is, conventional LiDAR systems are characterized by a synchronous processing arrangement in which Doppler shift (to calculate target velocity) and time delay (to calculate target distance) are measured at the same coherent processing interval (CPI). However, the inventors recognize that such a synchronous processing arrangement is arbitrary and that the asynchronous processing apparatus can be designed in which Doppler shift and time delay are measured at different CPIs. The inventors recognize that this asynchronous processing arrangement provides significant advantages, such as improved target velocity resolution when the CPI used to measure Doppler shift is longer than the CPI used to measure time delay.

[0042] In the following description, numerous specific details are set forth for purposes of explanation in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention may be practiced without these specific details. In other instances, well-known structures and devices are illustrated in block diagram form to avoid unnecessarily obscuring the invention.

[0043] 1. Overview of Phase Encoding Detection

[0044] Figure 1A This is a schematic diagram 110 illustrating an example of transmitting an optically phase-coded signal for distance measurement according to an embodiment. The horizontal axis 112 represents time from a start time at zero in arbitrary units. The left vertical axis 114a represents power during signal transmission in arbitrary units; and the right vertical axis 114b represents the phase of the transmitted signal in arbitrary units. For the simplest illustration of the phase-coded LiDAR technique, binary phase coding is illustrated. Trace 115 represents the power relative to the left axis 114a and is constant during signal transmission, decreasing to zero outside of transmission. The dashed trace 116 represents the phase of the signal relative to a continuous wave signal.

[0045] As can be seen, the trace is in phase with the carrier (phase = 0) used for a portion of the transmitted signal, and then changes Δφ (phase = Δφ) over short time intervals, repeatedly switching back and forth between the two phase values ​​on the transmitted signal, as shown in ellipsis 117. The shortest interval of constant phase is a parameter of the encoding, called the pulse duration τ, and is typically the duration of several cycles of the lowest frequency in the band. The reciprocal 1 / τ is the baud rate, where each baud represents one symbol. The number N of such constant-phase pulses during the transmission time is the number of symbols N, and represents the length of the code. In binary encoding, there are two phase values, and the phase of the shortest interval can be considered as 0 for one value and 1 for the other, so a symbol is one bit, and the baud rate is also called the bit rate. In polyphase encoding, there are multiple phase values. For example, four phase values ​​such as Δφ*{0, 1, 2, and 3}, where for Δφ = π / 2 (90 degrees), they are equal to {0, π / 2, π, and 3π / 2} respectively; and therefore, the four phase values ​​can represent 0, 1, 2, and 3 respectively. In this example, each symbol is two bits, and the bit rate is twice the baud rate.

[0046] Phase Shift Keying (PSK) is a digital modulation scheme that transmits data by changing the phase of a reference signal (carrier), such as... Figure 1A As shown. Modulation is applied by changing the sine and cosine inputs at precise times. In radio frequency (RF), PSK is widely used in wireless local area networks (LANs), RF identification (RFID), and Bluetooth communications. Alternatively, instead of operating relative to a constant reference wave, the transmitter can operate relative to itself. The phase change of a single transmitted waveform can be considered a symbol. In this system, the demodulator determines the phase change of the received signal, not the phase (relative to the reference wave) itself. Because this scheme depends on the difference between successive phases, it is called Differential Phase Shift Keying (DPSK). DPSK is much simpler to implement than ordinary PSK because it does not require the demodulator to have a copy of the reference signal to determine the precise phase of the received signal (it is an incoherent scheme).

[0047] For optical ranging applications, the carrier frequency is the optical frequency fc, and RF f0 is modulated onto the optical carrier. The number of symbols N and the duration τ are chosen to achieve the desired distance accuracy and resolution. The symbol pattern is chosen to distinguish it from other sources of compiled signal and noise. Therefore, a strong correlation between the transmitted and returned signals is a strong indication of the reflected or backscattered signal. The transmitted signal consists of one or more blocks of symbols, where each block is long enough to provide a strong correlation with the reflected or backscattered return signal even in the presence of noise. In the discussion below, it is assumed that the transmitted signal consists of M blocks of N symbols each, where M and N are non-negative integers.

[0048] Figure 1B This illustrates a series of binary numbers for distance measurement according to an embodiment. Figure 1A A schematic diagram 120 illustrates the example transmitted and returned optical signals. The horizontal axis 122 represents any unit of time after the start time at zero. The vertical axis 124a represents the amplitude of the transmitted optical signal at frequency fc+f0 in arbitrary units relative to zero. The vertical axis 124b represents the amplitude of the returned optical signal at frequency fc+f0 in arbitrary units relative to zero, and is offset from axis 124a to separate the trace. The trace 125 represents the transmitted signal of M*N binary symbols, having... Figure 1A The phase transitions shown produce a code starting with 00011010 and followed by an ellipsis. Trace 126 represents an idealized (noise-free) return signal scattered from a stationary object (and therefore, the return is without Doppler shift). The amplitude is reduced, but the code 00011010 is identifiable. Trace 127 represents an idealized (noise-free) return signal scattered from a moving object and therefore Doppler shifted. The return signal is not at the appropriate optical frequency f. c At +f0, and not well detected in the expected frequency band, the amplitude is significantly reduced.

[0049] Using the Doppler effect given in Equation 1, the observed returned frequency f' is different from the returned correct frequency f = f c +f0 is different.

[0050]

[0051] Where c is the speed of light in the medium. Note that if the observer and the source move in the same direction with the same velocity along the vector between them, then the two frequencies are the same. The difference between the two frequencies, Δf = f' - f, is the Doppler frequency shift Δf. D It will cause problems for distance measurement, as given by Equation 2.

[0052]

[0053] Note that the magnitude of the error increases with the signal frequency f. Also note that for a static LiDAR system (v0 = 0), for an object moving at 10 meters per second (v0 = 10) and visible light at a frequency of approximately 500 THz, the error is approximately 16 MHz (1 MHz = 10 THz). 6 The frequency is on the order of Hertz (Hz, 1 Hz = 1 cycle per second). In the various embodiments described below, Doppler frequency shift error is detected and used to process data for distance calculation.

[0054] Figure 1CThis is a schematic diagram 130 illustrating an example cross-correlation of a reference signal with two return signals according to an embodiment. In phase-compiled ranging, the arrival of a phase-compiled reflection can be detected in the return by cross-correlation of the transmitted signal or other reference signal with the return signal. In practice, this can be implemented by using heterodyne detection to cross-correlate the code of the RF signal with an electrical signal from an optical detector and thus downmixing it back to the RF band. The horizontal axis 132 indicates the lag time applied to the compiled signal before cross-correlation calculation with the return signal in arbitrary units. The vertical axis 134 represents the magnitude of the cross-correlation calculation. The cross-correlation for any lag is calculated by convolving the two traces (i.e., by multiplying corresponding values ​​in the two traces and summing over all points in the traces) and then repeating for each time lag. Alternatively, the cross-correlation can be implemented by multiplying the Fourier transforms of each of the two traces, followed by the inverse Fourier transform. Efficient hardware and software implementations for the Fast Fourier Transform (FFT) are widely used for both forward and inverse Fourier transforms. Below, more precise mathematical expressions for cross-correlation are provided for some example implementations.

[0055] Note that the cross-correlation calculation is performed using analog or digital electrical signals after the amplitude and phase of the returned signal are detected at the optical detector. To shift the signal at the optical detector into an RF frequency range that can be easily digitized, the optically returned signal is optically mixed with a reference signal before striking the detector. A phase-encoded copy of the transmitted optical signal can be used as the reference signal, but it is also possible, and generally preferred, to use an optical signal at the continuous-wave carrier frequency output from the laser, capturing both the amplitude and phase of the electrical signal output from the detector.

[0056] Trace 136 represents the cross-correlation with the idealized (noise-free) return signal, which is reflected from a stationary object (and therefore the return signal is not Doppler shifted). The peak appears at time Δt after the start of the transmitted signal. This indicates that the return signal includes a version compiled from the transmitted phase starting at time Δt. The distance (range) L to the reflecting (or backscattered) object is calculated based on the speed of light c in the medium according to the two-way travel time delay, as given by Equation 3.

[0057] L=c*Δt / 2 (3)

[0058] The dashed line 137 represents the cross-correlation with the idealized (noise-free) return signal scattered from the moving object (and therefore the return is Doppler-shifted). The return signal does not include phase encoding within the appropriate frequency bin, the correlation remains low for all time lags, and the peaks are not easily detected. Therefore, Δt is not easily determined, and the distance L is not easily generated.

[0059] According to the various embodiments described in more detail below, the Doppler frequency shift is determined in the electrical processing of the returned signal; and the Doppler frequency shift is used to correct cross-correlation calculations. Therefore, peaks are easier to find, and distances can be more easily determined. Figure 1D This is a schematic diagram 140 illustrating an example spectrum of the transmitted signal and an example spectrum of the Doppler-shifted return signal according to an embodiment. The horizontal axis 142 represents the RF frequency shift relative to the optical carrier fc in arbitrary units. The vertical axis 144a represents the amplitude of a specific narrow frequency range (also called spectral density) in arbitrary units relative to zero. The vertical axis 144b represents the spectral density in arbitrary units relative to zero and is offset from axis 144a to separate the traces. Trace 145 represents the transmitted signal; and the peak occurs at the appropriate RF f0. Trace 146 represents the idealized (noise-free) return signal, which is backscattered from a moving object and is therefore Doppler-shifted. The return signal has no peak at the appropriate RF f0; instead, it is blue-shifted by Δf. D to the shift frequency f S .

[0060] In some Doppler compensation implementations, it is not as... Figure 1D The method shown involves acquiring the spectra of the transmitted and returned signals, searching for peaks in each, and then subtracting the frequency of the corresponding peak to find Δf. D Instead, it is more efficient to obtain the cross spectrum of the in-phase and quadrature components of the downmixed return signal in the RF band. Figure 1E This is a schematic diagram 150 illustrating an example cross spectrum according to an embodiment. The horizontal axis 152 represents the shift relative to the reference spectrum in arbitrary units; and the vertical axis 154 represents the amplitude of the cross spectrum in arbitrary units relative to zero. The trace 155 represents a cross spectrum with an idealized (noise-free) return signal, which is moved by an object toward the LIDAR system (…). Figure 1D Δf in D1 =Δf D Blue shift) and the second object moving away from the LiDAR system (Δf D2 The redshift is generated. When one of the components, the blueshift Δf, is... D1 A peak appears at a certain time; and, when one of the components redshifts by Δf D2 Another peak appears at this time. Therefore, the Doppler frequency shift is determined. These shifts can be used to determine the approach velocity of objects near LiDAR, which may be crucial for collision avoidance applications.

[0061] As described in more detail below, the Doppler shift detected in the cross spectrum is used to correct the cross-correlation, making peak 135 apparent in the Doppler-compensated Doppler shift return with hysteresis Δt, and the distance L can be determined. The information required to determine and compensate for the Doppler shift is either not collected or not used in existing phase-coded LiDAR systems.

[0062] 2. Overview of Optical Inspection Hardware

[0063] To describe how to implement the phase-coded detection method, some general and specific hardware methods are described. Figure 2 This is a block diagram illustrating example components of a high-resolution LiDAR system according to an embodiment. A laser source 212 emits a carrier wave 201, which is phase-modulated in a phase modulator 282 to generate a phase-compiled optical signal 203 having a symbol length M*N and a duration D = M*N*τ. A beam splitter 216 splits the optical signal into a target beam 205 (also referred to herein as the transmitted signal) having most of the energy of beam 203, and a reference beam 207a having very little energy, but sufficient to produce good mixing with the return beam 291 scattered from an object (not shown). In some embodiments, the beam splitter 216 is positioned upstream of the phase modulator 282. The reference beam 207a passes through a reference path 220 and is directed as a reference beam 207b to one or more detectors. In some embodiments, the reference path 220 introduces a known delay sufficient to allow the reference beam 207b to reach the detector array 230 along with the scattered light. In some implementations, the reference beam 207b is referred to as a local oscillator (LO) signal, referring to an older method in which the reference beam 207b is locally generated from separate oscillators. In various implementations, methods ranging from less flexible to more flexible involve the following steps to ensure the reference arrives along with the scattered or reflected field: 1) placing a mirror in the scene to reflect a portion of the transmitted beam back to the detector array, resulting in a good path length match; 2) using fiber delay to closely match the path length and propagating the reference beam using optics near the detector array, such as... Figure 2 As suggested in the document, it may or may not have path length adjustment to compensate for phase differences observed or anticipated for a specific distance; or, 3) use a time delay of frequency-shifting devices (acousto-optic modulators) or local oscillator waveform modulation to generate separate modulation to compensate for path length mismatch; or some combination thereof. In some implementations, the objects are close enough and the transmission duration is long enough that the return signal fully overlaps with the reference signal without delay.

[0064] The detector array is a single paired or unpaired detector, or a 1D or 2D array of paired or unpaired detectors arranged in a plane substantially perpendicular to the return beam 291 from the object. The reference beam 207b and the return beam 291 are combined in zero or more optical mixers to produce an optical signal with the characteristics to be properly detected. The acquisition system 240 records the phase or amplitude, or some combination thereof, of the interference modes for each detector multiple times during the signal duration D. The number of time samples for each signal duration affects the down-range extent. The number is typically chosen based on practical considerations such as the number of symbols per signal, the signal repetition rate, and the available camera frame rate. The frame rate is the sampling bandwidth, often referred to as the "digitizer frequency." The only fundamental limitation to the range is the coherence length of the laser and the length of the linear frequency modulation or unique code prior to its repetition (for explicit ranging). This is enabled because any digital record of the returned bits can be cross-correlated with any portion of the transmitted bits from the previously transmitted history. The acquired data can be used for processing by system 250, as described in the reference below. Figure 12 The computer system described, or the reference below. Figure 13 The chipset described herein. The Doppler compensation module 270 determines the magnitude of the Doppler frequency shift and the correction range based thereon, as well as any other corrections described herein. Any known device or system can be used to implement the laser source 212, phase modulator 282, beam splitter 216, reference path 220, optical mixer 284, detector array 230, or acquisition system 240. Optical couplings used for diffuse or focus on a target or for focusing across the pupil plane are not depicted. As used herein, an optical coupler is any component used to influence the propagation of light in spatial coordinates to guide light from one component to another, such as vacuum, air, glass, crystal, mirror, lens, optical circulator, beam splitter, phase plate, polarizer, optical fiber, optical mixer, etc., individually or in some combination.

[0065] 3. Phase-encoded optical detection

[0066] In some implementations, an electro-optic modulator provides modulation. The system is configured to generate phase codes of length M*N and symbol duration τ, suitable for the desired forward range resolution, as described in more detail below for various implementations. In implementations, the phase code comprises a plurality of M blocks, each having a duration of N*τ, or phase code duration. For example, in 3D imaging applications, the total number of pulses M*N is in the range of about 500 to about 4000. Since processing is typically performed in the digital domain, it is advantageous to choose M*N as a power of 2, for example, in the range from 512 to 4096. When no averaging is performed, M is 1. If random noise components are present, it is advantageous for M to be about 10. As a result, for M=1, N is in the range from 512 to 4096, and for M=10, N is in the range from about 50 to about 400. For baud rates from 500 Mbps to 1 Gbps, the duration of these codes is between about 500 ns and 8 microseconds. Note that under these conditions, the distance window can be extended to several kilometers, and the Doppler resolution can also be quite high (depending on the duration of the transmitted signal). Although in Figure 2 In this document, for illustrative purposes, processes, devices, and data structures are described as integral blocks in a particular arrangement. However, in other embodiments, one or more processes or data structures, or portions thereof, may be arranged differently on the same or different hosts, in one or more databases, or may be omitted, or one or more different processes or data structures may be included on the same or different hosts. For example, beam splitter 216 and reference path 220 include zero or more optical couplers.

[0067] 3.1 Doppler Compensation LiDAR

[0068] Figure 3A This is a block diagram illustrating example components of a phase-coded LiDAR system 300a. Although object 390 is depicted to illustrate the operation of system 300a, object 390 is not part of system 300a. The system includes a laser source 310, a beam splitter 312, a phase modulator 320, a polarization beam splitter 322, an optical mixer 360, a photodetector 330 (also referred to herein as "optical detector 330"), and a processing system 350, which includes a digital code module 372 and a Doppler compensation module 370. Optical signals are indicated by thick arrows, and electrical signals by thin arrows.

[0069] In electrical engineering, a sine curve with phase modulation (corresponding to angular modulation between the real and imaginary parts of the mathematical function exp(iωt)) can be decomposed into two amplitude-modulated sine curves offset by a quarter period (π / 2 radians) in phase, or synthesized from these two amplitude-modulated sine curves. All three functions have the same frequency. The amplitude-modulated sine curves are referred to as the in-phase component (I) at phase 0 and the quadrature component (Q) at phase π / 2. Laser 310 generates an optical signal at the carrier frequency fc. The laser optical signal L is mathematically represented by Equation 4.

[0070] L=I0 exp(iωt) (4)

[0071] Where I0 is the intensity of the laser output, and exp() is an exponential function such that exp(x) = e x Let i be an imaginary number with the property of having the square root of -1, t be time, and ω = 2πfc be the angular frequency corresponding to the optical carrier frequency fc. Mathematically, this expression has a real part of I. 0R cos(ωt) and imaginary part = I 0I sin(ωt), where I 0R It is the real part of the intensity (in phase), and I 0I It is the imaginary part. The phase of the oscillation is given by the angle between the real and imaginary parts. Therefore, L = I 0R cos(ωt)+iI 0I sin(ωt), and I0 is the root of the sum of the squares of the real and imaginary parts, I0 2 =I 0R 2 +I 0I 2 The beam splitter 312 directs a small portion of the signal strength to be used as a reference signal (called the local oscillator) LO, as given by Equation 5.

[0072] LO = A LO exp(iωt)=A R cos(ωt) + iA1 sin(ωt). (5a)

[0073] Where A is a constant representing the intensity effect of beam splitter 312. Electric field E LO Therefore, it can be written as equation 5b.

[0074] E LO =A LO e iωt (5b)

[0075] When the reference signal (LO) is an unmodulated laser signal, the entire signal is in phase and its imaginary part is zero.

[0076] LO = A cos(ωt). (5c)

[0077] The digital code module 372 in the processing system 350 transmits an electrical signal representing a digital code, denoted as B(t), of the sign applied as a function of the phase change on the optical carrier, where B(t) switches between 0 and π / 2 as a function of t. The phase modulator 320 imposes the phase change on the optical carrier by taking digital lines from the field-programmable gate array (FPGA), amplifying them, and driving the EO phase modulator. The transmitted optical signal T is then given by Equation 6.

[0078] T=C eip(i[ωt+B(t)]) (6)

[0079] Where C is a constant that takes into account the reduction in I0 caused by the division of A and any amplification or further reduction applied by the phase modulator 320.

[0080] Any phase modulator can be used as modulator 320. For example, an electro-optic modulator (EOM) comprising a crystal such as lithium niobate, whose refractive index is a function of the local electric field strength, can be used. This means that if lithium niobate is exposed to an electric field, light will pass through it more slowly. However, the phase of the light leaving the crystal is proportional to the length of time it takes for the light to pass through the crystal. Therefore, the phase of the laser leaving the EOM can be controlled by changing the electric field in the crystal according to a digital code provided by digital code module 372. The phase change results in a broadband frequency signal, where the bandwidth B is approximately equal to the baud rate 1 / τ.

[0081] The phase-encoded optical signal output from phase modulator 320 is transmitted via some optical coupler (such as polarization beam splitter (PBS) 322 or other circulator optics), and then scattered by any object 390 in the beam carrying the transmitted signal. For example, it has been found that fiber-coupled polarization beam splitter combiners provide better isolation between ports than fiber-based circulators as such optical components. This is important because signals not well isolated between transmission and reception will appear as undesirable large peaks in the range profile. Therefore, the transmitted signal is injected into port 1, transmitted from port 2, and the backscattered return signal is received in port 2 and exits from port 3. Some targets (e.g., metallic targets) maintain the polarization of the beam, while some targets (e.g., diffuse targets) depolarize the returning beam. In some embodiments, a quarter-wave plate is included in the transmitting optics to appropriately compensate for undepolarized targets.

[0082] The return signal 324 is guided by an optical coupler (e.g., PBS 322) to an optical mixer 360, where the return optical signal 324 is mixed with a reference optical signal (LO) 314 given by Equation 5, and the return signal R from the transmitted beam intercepted by the k-th object is given by Equation 7a.

[0083] R k =A k exp(i[(ω+ω Dk (t+Δt) k )+B(t+Δt k (7a)

[0084] Where A k It is a constant Δt that takes into account the intensity loss due to propagation to and from object 390 and scattering at the k-th object 390. k ω is the bidirectional travel time between the LIDAR system and the k-th object 390, and ω Dk =2πΔf D It is the angular frequency of the Doppler shift of the k-th object (here referred to as the Doppler shift for convenience). Then, the electric field of the return signal ER, summed over all targets, is given by Equation 7b.

[0085]

[0086] The coincident signals at optical mixer 360 (e.g., return optical signals 324 and LO 314) generate a mixed optical signal 362 having a beat frequency related to the difference in frequency, phase, and amplitude of the two optical signals being mixed, and produce an output depending on the function of optical mixer 360. As used herein, undermixing refers to optical heterodyne detection, which is an implementation of the heterodyne detection principle using a nonlinear optical process. In optical heterodyne detection, referred to herein as “undermixing,” an optical signal of interest at a certain optical frequency is nonlinearly mixed with a reference “local oscillator” (LO) set at a similar frequency. The desired result is a difference frequency, which carries information (amplitude, phase, and frequency modulation) of the original optical frequency signal, but conveniently oscillates in the RF band at a lower, more easily processed frequency (here referred to as the beat frequency). In some implementations, this beat frequency is in the RF band that can be output as an electrical signal 332 from optical detector 330, such as an electrical analog signal that can be easily digitized by an RF analog-to-digital converter (ADC). Electrical signal 332 is input to processing system 350 and, together with digital code from digital code module 372, is used by Doppler compensation module 370 to determine cross-correlation and distance, and in some embodiments, to determine velocity and Doppler frequency shift.

[0087] In some implementations, the raw signal is processed to find the Doppler peak, and the frequency ω is used. D This is used to correct the correlation calculation and determine the correct distance. In other implementations, it has been found that if the optical mixer and processing are configured to determine the in-phase and quadrature components, and this separation is used to first estimate ω... D And then use ω D It would be advantageous to correct the cross-correlation calculation to derive Δt. ω D The value is also used to render the object's speed, and the first time period is selected to adjust ω. D The resolution and velocity of the object. Using Equation 3 above, the value of Δt is used to determine and represent the distance to the object. Separating the I and Q signals using an optical mixer allows for clear determination of the sign of the Doppler shift.

[0088] This diagram illustrates an example hardware implementation for coherent detection of in-phase and quadrature (I / Q) signals of a transmitted signal that supports phase coding. The advantage of this approach is its very low cost yet high bandwidth, low waveform generation requirements (binary digital or polyphase digital codes), and minimal modulation requirements (a single electro-optic phase modulator). A 90-degree optical mixer allows for I / Q detection of the optically mixed signal on both channels, followed by digitization. This system allows for a very flexible, "software-defined" measurement architecture.

[0089] Figure 3B This is a block diagram illustrating example components of a Doppler-compensated phase-coded LiDAR system 300b according to an embodiment. This embodiment uses binary phase coding, where the two phases are separated by π / 2, but with optical separation instead of electrical separation, having in-phase and quadrature components. Although object 390 is depicted to illustrate the operation of system 300a, object 390 is not part of system 300a. The system includes a laser source 310, a beam splitter 312, a phase modulator 320, a polarization beam splitter 322, and alternatively... Figure 3A The general-purpose optical mixer 360's 90-degree mixing mixer 361, replacing Figure 3A The photodetector 330 includes a balanced photodetector 331, and a processing system 350, which includes a digital code module 372 and a Doppler compensation module 371. Optical signals are represented by thick arrows, and electrical signals by thin arrows. A laser 310 generates an optical signal at the optical carrier frequency fc. A beam splitter 312 directs a small portion of the signal power to be used as a reference signal (called a local oscillator) LO 314. The digital code module 372 in the processing system 350 transmits an electrical signal representing a digital code (e.g., M blocks, each block having a phase code duration N*τ) to be applied to the optical carrier as a phase change. As described above, a phase modulator 320 applies a phase change to the optical carrier.

[0090] The phase-encoded optical signal output from phase modulator 320 is transmitted through some optical couplers (such as polarization beam splitter (PBS) 322), after which it is scattered by any object 390 intercepted by the beam carrying the transmitted signal. The return signal 324 is guided by an optical coupler such as PBS 322 to a 90-degree hybrid optical mixer 361, where the return optical signal 324 is mixed with a reference optical signal (LO) 314 given by equation 5b. The return signal R is given by equation 7a. The hybrid mixer outputs four optical signals, referred to as I+, I-, Q+, and Q-, which combine the in-phase component of LO with the return signal R (referred to as RI) and the quadrature component of the return signal R (referred to as R0). Q Combinations, as defined in equations 8a to 8d.

[0091] I+=LO+R1 (8a)

[0092] I-=LO-R I (8b)

[0093] Q+=LO+R Q (8c)

[0094] Q-=LO-R Q (8d)

[0095] Where R I It is the in-phase coherent cross term of the AC component of the returned signal R, while R Q It is the 90-degree out-of-phase coherent cross term of the AC component of the returned signal R. For example, the electric field of the above relationship can be expressed based on equations 5b and 7b above and equations 8e to 8g below to generate equations 8h to 8k.

[0096] LO = |E LO | 2 (8e)

[0097]

[0098]

[0099] Where * denotes the complex conjugate of a complex number, Imag() is a function that returns the imaginary part of a complex number, and Real() is a function that returns the real part of a complex number. (AC) All optical frequency components of the signal are eliminated, leaving only the RF "beat frequency" of the LO and the RF component of the returned signal—in this case, the Doppler frequency shift and code function. Item | E LO | 2 and |E R | 2It is a constant (direct current, DC) term. The latter is negligible relative to the former; therefore, the latter term is ignored in the combinations expressed in equations 8h to 8k, as a specific form of equations 8a to 8d.

[0100]

[0101]

[0102]

[0103]

[0104] According to equations 9a and 9b, the two in-phase components I+ and I- are combined at the balanced detector pair to generate an RF signal I on channel 1 (Ch1), and the two quadrature components Q+ and Q- are combined at the second balanced detector pair to generate an RF signal Q on channel 2 (Ch2).

[0105] I = I + -I - (9a)

[0106] Q = Q+ - Q- (9b)

[0107] Using a balanced detector (a pair of balanced optical detectors) offers the advantage of eliminating common-mode noise, providing reliable measurements with a high signal-to-noise ratio (SNR). In some implementations, this common-mode noise is negligible or otherwise unimportant; therefore, a simple optical detector or an unbalanced pair is used instead of a balanced pair.

[0108] In some implementations, the LO signal alternates between in-phase and quadrature versions of the transmitted signal, such that electrical signals I and Q are measured at close but different times of equal duration.

[0109] Then, the Doppler compensation module 371 uses signals I and Q to determine one or more Doppler frequency shifts ω with corresponding velocities over a time period of at least one code block duration. D In some implementations, assuming that multiple blocks are sampling the same object or are expected to sample the same object, the resolution of the Doppler shift (and thus the velocity resolution) is increased by extending the first duration to a multiple of the duration of a code block, as explained in more detail below.

[0110] ω from digit code module 372 DThe values ​​of B(t) and I, along with signals I and Q, are then used to generate corrected correlation traces over corresponding time periods (e.g., a second duration having the duration of at least one code block), where peaks represent one or more Δt values ​​for each of the one or more velocities. When multiple velocities are detected, each velocity is associated with a peak in one or more corresponding correlation traces. In some implementations, this is accomplished through an overlap process to determine which current velocity / position pair is most likely associated with a previous similar velocity / position pair. The one or more Δt values ​​are then used, using Equation 3 described above, to determine one or more distances. To increase distance resolution, it is desirable to perform this calculation over the shortest possible time period, e.g., the second duration is equal to the duration of a code block.

[0111] Therefore, the first duration and the second duration are usually different. For increased Doppler shift resolution and increased range resolution, it is advantageous for the first duration to be longer than the second duration. This can be achieved, for example, by storing several previously returned blocks in a storage buffer to extend the first duration.

[0112] Advantageously, a frequency domain representation of the code used for correlation is prepared at the beginning and repeated for the frequency domain representation of the code used at each measurement point in the scan; thus, this is done in some embodiments. A long code with a duration D = (M*N)*τ is encoded onto the transmitted light, and the data acquisition electronics collect the return signal of the same duration. The Dyna and the signal are both divided into M shorter blocks of length N and phase code duration N*τ so that correlation can be performed several times on the same data stream, and the results are averaged to improve the signal-to-noise ratio (SNR). Each block of N symbols and phase code duration N*τ is distinct from the different blocks of N symbols, so each block is measured independently. Therefore, averaging reduces noise in the return signal. The phase separation of the input I / Q signals is π / 2. In some embodiments, further averaging is performed at several illumination points to remove the effects of reflections from purely internal optics, as described in previous work.

[0113] 3.2. Optical Detection Methods

[0114] The proposed method increases resolution or signal-to-noise ratio, or both, to compute a cross spectrum using the phase difference of I / Q signals (in electrical or optical signals), which provides a clear peak at the Doppler frequency. The method also utilizes the phase difference of the I / Q signals to construct a complex signal for correlation to determine the distance. Doppler compensation is accomplished by first employing an FFT of the complex return signal and then shifting the FFT value within an array of frequency intervals. The corrected signal can be recovered by applying an inverse FFT to the shifted FFT, but this is not necessary because in some implementations, the shifted FFT is used directly for correlation with the code FFT. In other implementations, the complex return signal is multiplied by a complex exponent formed by the Doppler frequency measured in the cross spectrum, and the FFT of the corrected signal is used for correlation with the code. In some implementations, a finite impulse response (FIR) filter is used to determine the correlation. After calculating the correlation (also referred to herein as the distance profile) for each code / signal block, the results are averaged over M blocks, and the distance to the target is calculated based on the time delay of the peak in the average distance profile. If there is more than one peak in the distance profile, the method will record the distances to multiple targets. The proposed method utilizes asynchronous processing of Doppler shift and distance to the target at different time intervals, thereby optimizing the resolution of the Doppler shift and the velocity of the object.

[0115] Figure 4A This is a flowchart illustrating an example method 400 for determining and compensating for the Doppler effect on distance in a phase-coded LIDAR system using Doppler correction, according to an embodiment. Although the steps are for illustrative purposes... Figure 4A and Figure 4B The process is described as a series of steps in a specific order; however, in other embodiments, one or more steps or portions thereof are performed in a different order, or overlap, are performed in series or in parallel in time, or are omitted, or one or more additional steps are added, or the method is changed in some combination of ways. In some embodiments, steps 403, as well as steps 410 to 433, and / or steps 451 to 461, are performed by processing system 350. For example, the FFT for determining the digit code in step 403 and all steps in steps 410 to 433, and / or steps 451 to 461 are performed by... Figure 3A Doppler compensation module 370 or Figure 3B Module 371 in the middle is used.

[0116] In step 401, the transceiver, such as a LiDAR system, is configured to transmit a phase-coded optical signal based on an input phase code sequence. A portion (e.g., 1% to 10%) of the unmodulated input optical signal from the laser or the phase-coded transmitted signal is also directed to a reference optical path. The transceiver is also configured to receive backscattered optical signals from any external object illuminated by the transmitted signal. In some embodiments, step 401 also includes hardware-configuring other optical components to provide the functionality of one or more of the following steps, such as, for example… Figure 3A or Figure 3B Or an equivalent is shown. Note that the transmitted signal does not have to be a beam. A diverging signal will certainly have many different distances and Doppler values ​​within a single range profile; however, it does not provide lateral range resolution within the illuminated point. Nevertheless, it is advantageous to use a narrow beam that provides the inherent sparsity of point-by-point scanning to provide lateral range resolution useful for object identification.

[0117] In step 403, a code consisting of M*N symbol sequences is generated for ranging, representing M blocks of N symbols, where there is no repetition between the M blocks. In some embodiments, the Fourier transform of the RF signal with this phase encoding is also determined during step 403, since the transform can be reused in step 423 as described below, and it is advantageous that the Fourier transform does not need to be computed separately for each transmission. For example, a complex digital signal (real and imaginary parts) with an angular RF frequency ω and a phase π / 2 is generated based on the generated code, and a complex digital fast Fourier transform (FFT) is computed on this complex digital signal. The resulting complex FFT function is prepared for the operation in step 423 by employing the complex conjugate of the complex signal. For example, for each of the M blocks of the code, the complex conjugate Code of the complex FFT is represented by Equation 10. FFT .

[0118] Code FFT =conj(FFT(exp(iBt)) (10)

[0119] Here, conj() represents the complex conjugate operation, which is conj(x+iy) = x-iy. This complex FFT is stored, for example, on a computer-readable medium for subsequent use during step 423, as described below.

[0120] In step 405, the first portion of the laser output, represented by Equation 4, is phase-encoded using the code received from the digital code module 372 to generate a transmitted phase-encoded signal as shown in Equation 6, and this first portion is directed to a point in the scene where an object or part of an object may or may not be present. Additionally, in step 405, a second portion of the laser output is directed along a reference path as a reference signal, as shown in Equation 5a or Equation 5b, also referred to as a local oscillator (LO) signal.

[0121] In step 407, as shown in Equation 7, there is any travel time delay Δ t and Doppler frequency shift ω D The backscattered return signal R is mixed with a reference signal LO as shown in Equation 5a or Equation 5b to output one or more mixed optical signals 362. The mixed signal informs the in-phase and quadrature components. For example, in Figure 3B In the illustrated embodiment, the mixed optical signal 362 includes four optical signals, I+, I-, Q+, and Q-, that notify the in-phase and quadrature components, as defined in equations 8a to 8d. In other embodiments, different optical mixers are used. For example, in some embodiments, a 3×3 coupler is used instead of a 90-degree optical mixer to still support I / Q detection.

[0122] In step 408, the mixed optical signal is directed to one or more optical detectors and detected at one or more optical detectors to convert the optical signal into one or more corresponding electrical signals. For example, in Figure 3B In the illustrated embodiment, the detector generates two electrical signals. One signal on one channel (Ch 1) represents the in-phase component I of the lower mixing, as given by Equation 9a; the other signal on the other channel (CH 2) represents the quadrature component Q of the lower mixing, as given by Equation 9b. The complex lower-mixed signal S is calculated based on the two electrical signals, as given by Equation 11.

[0123] S=I+iQ (11a)

[0124] Note that signals S, I, and Q are functions of time t with a duration of at least D = M * N * τ.

[0125] In some implementations, averaging is performed on several different return signals S(t) to remove pseudo-copies of the phase-coded signal generated at internal optics such as PBS 322 along the return signal path. Such pseudo-copies can reduce the correlation with the actual return from an external object, thus masking the actual return as almost undetectable. If averaging is performed on P different illumination points and returns such that a single object is not in all of those illumination points, the averaging is dominated by pseudo-copies of the code generated by the internal optics. The pseudo-copies of this code can then be removed from the return signal to leave only the actual return in the corrected complex electrical signal S(t). P is a sufficiently large number to ensure that the same object is not illuminated at all points. Values ​​as low as P = 100 are computationally advantageous for graphics processing unit (GPU) implementations; while values ​​up to P = 1000 are preferred and suitable for field-programmable gate array (FPGA) implementations. In the example implementation, P is approximately 100. In other implementations, depending on the application, P can range from approximately 10 to approximately 5000. Figure 11 is a block diagram illustrating an example multi-point averaging method for removing returned samples from internal optics according to an embodiment. This correction is performed in steps 409 and 410.

[0126] In step 409, it is determined whether P responses have been received. If not, control returns to step 405 to illuminate another point. If yes, control returns to step 410. In step 410, the average signal S is calculated according to equation 11b. S Sp(t), where each received signal of duration D is designated as Sp(t).

[0127]

[0128] This average signal is used to correct each received signal Sp(t) to generate the corrected signal S. pC (t), the correction signal S pC (t) is used as the received signal S(t) in subsequent steps, as given by equation (11c).

[0129] S(t)=S pC (t)=S p (t)-S S (t) (11c)

[0130] In some implementations, the internal optics are calibrated once under controlled conditions to produce S. SA fixed value for S(t) is stored for multiple subsequent deployments of the system. Therefore, step 410 only involves applying equation 11c. In some embodiments, the pseudo-copy of the code generated by the internal optics is small enough, or the associated distance is sufficiently different from the distance to the external object, so that steps 409 and 410 can be omitted. Therefore, in some embodiments, steps 409 and 410 are omitted, and control proceeds directly from step 408 to step 411, using S(t) from step 408 instead of equation 11c in step 410.

[0131] In some implementations, during step 410, the average signal S is used as the basis. S Additional corrections are applied to the electrical signal S(t). For example, as described in more detail in section 4.4, the signal S at different points... p The evolution of (t) or the average signal S at each of the p points. S The phase and frequency drift of the laser are detected in the evolution of (t). This observed drift is used to formulate a correction that digitally compensates for laser linewidth issues caused by hardware or other noise sources. In another example, described in more detail in Example Section 4.5, a drift with time evolution at a shorter scale (e.g., at the scale of each block of N compiler symbols) is used to compensate for the reduction in signal-to-noise ratio (SNR) due to coherent broadening in the Doppler frequency domain.

[0132] In step 411, the cross spectrum is used to detect the Doppler shift. The following explanation is provided for interpretive purposes; however, the characteristics and utility of various techniques are not limited by the accuracy or completeness of this explanation. The frequency contents of I and Q contain both the Doppler (sine) and code (square wave) components. For the Doppler component, since it is sinusoidal, I is expected to lag or advance by 90 degrees compared to Q. The lag or advance depends on the sign of the Doppler shift. The code component does not plot this effect—it indicates whether the I and Q levels of the time-based return bits are in-phase or 180-degree out-of-phase. The operation within parentheses of the XS operation calculates the complex phasor difference between I and Q at a given frequency. If there is a 90-degree phase difference between I and Q at a given frequency (as in the case of the Doppler component), this will be reflected in the imaginary part of the result. Conversely, the code frequency contents will not appear in the imaginary part of the result because, as stated above, for the selected binary code, the I and Q aspects of the code are either in-phase or 180-degree out-of-phase, so the complex phasor difference at each frequency is always real. The cross-spectral operation XS() can be viewed as a way to reveal only those aspects of the signal spectrum relevant to Doppler, where the code is lost. This makes it easier to find the Doppler frequency content. Conversely, in the regular spectrum of the returned signal, the code frequency content may obscure the Doppler frequency content expected for good Doppler estimation / correction.

[0133] For example, as given in Equation 12, calculate the cross spectrum of S.

[0134] XS(S)=FFF(I)*conj[FFT(Q)] (12)

[0135] XS(S) obtained from Equation 12 is a complex-valued array. The peaks in this cross spectrum represent one or more Doppler frequency shifts ω in the returned signal. D Note that ω D =2πΔf D Any peak detection method can be used to automatically determine peaks in a cross spectrum XS(S). Typically, identifying large positive or negative peaks in the imaginary part of the cross spectrum will reveal information about the Doppler shift. However, in some special cases, the real part may also reveal such information. An example of this is the presence of multiple distance returns with similar Doppler values. The magnitude of the increase in the real part can represent this situation. In some implementations, the cross spectrum operation is performed individually on each data block and averaged over M blocks. These Doppler shifts and corresponding relative velocities are stored for further use, for example, on one or more computer-readable media. As described further in detail below, the power spectrum is also useful for identifying Doppler shifts and obtaining phase.

[0136] In some implementations, the Doppler shift is calculated over several blocks to increase frequency resolution (and thus velocity resolution). The resolution of velocity measurements is essentially limited by the coherent processing interval (CPI) (e.g., the duration of a symbol block with a duration equal to N*τ). The CPI limits the frequency resolution of the measurement to 1 / CPI and ultimately limits the Doppler frequency and the resolution of the corresponding velocity measurements in the LiDAR system. The asynchronous processing of the measurement signal for Doppler and range is flexible. If a larger Doppler shift resolution (and therefore a larger velocity resolution) is desired, it is possible to buffer the duration of time-domain data longer than the duration of a block of phase-compiled waveform. This time-domain data segment can then be analyzed using cross-spectrum or power spectrum to solve for the target velocity with finer velocity resolution.

[0137] The increased duration comes with some computational cost, as cross-spectrum computation increases with the number of samples in the signal. However, the sampling rate (samples per second) determines the highest definite frequency (Nyquist frequency = sampling rate / 2). The Nyquist frequency typically corresponds to a Doppler shift velocity much greater than any expected velocity. Therefore, by downsampling to a lower sampling rate (e.g., by averaging several consecutive samples before computing the cross-spectrum), computational costs (e.g., shorter FFTs and shorter peak searches) can be reduced without losing meaningful velocity measurement space. These concepts are discussed in... Figures 5A to 5C As shown in the image.

[0138] Figure 5A This is a block diagram illustrating an example of multiple time blocks according to an embodiment and a first duration longer than each time block, wherein each time block is Figure 1B The duration of a block of phase code. In some implementations, multiple blocks with a first duration are processed together and downsampled to determine a high-resolution Doppler shift without loss of the expected Doppler shift. Each block of N symbols with a shorter second duration is processed individually for the distance value using the corresponding Doppler shift measurement; and the M blocks are averaged to increase the signal-to-noise ratio of the distance to a single point. The first duration used to calculate the Doppler shift can be shorter or longer than the M blocks averaged for each point. Figure 5B This is a graph illustrating the relationship between an example power spectrum according to an embodiment and frequencies calculated for each time block and for a first duration across multiple time blocks. Two returns with different Doppler shifts are decomposed in a solid trace where the first duration equals four code blocks (4 x CPI). The two returns are not distinguished in a dashed trace using a duration equal to one block (duration equal to CPI). In various embodiments, the consecutive time intervals used for calculating the first duration of the Doppler shift can overlap, be continuous, or be discontinuous. Figure 5C This illustrates an embodiment. Figure 5A A block diagram illustrating an example of multiple time blocks and multiple time periods of a first duration, wherein consecutive time periods overlap. An example implementation is described in more detail in section 4.1.

[0139] These methods result in asymmetric power spectra or cross spectra, allowing the sign of the Doppler shift to be discerned by the position of the beat peaks from the residual carrier. In some implementations, the electrical signals are real-valued; for example, there are no separate electrical signals for the real and complex parts of the return, or for the in-phase and quadrature parts of the return signal. In such implementations, the Doppler shift is still determined by the cross spectrum of two identical time series, which is equivalent to an automatic spectrum, simply referred to as the spectrum. Thus, as used herein, the spectrum refers to the cross spectrum when the two time series are identical. Therefore, if only a single phase of the optical field is measured in the time domain, the input time-domain data is real, and thus, after the FFT, the power spectrum or cross spectrum is symmetric about the DC (f=0) frequency interval, and the negative half of the spectrum is exactly the complex conjugate of the positive half of the spectrum. Therefore, if a Doppler shift exists, two identical amplitude peaks symmetric about the 0 Doppler shift are observed, and the sign of the shift is unknown.

[0140] In step 413, for example, a complex FFT function FFT(S) implemented in hardware or software is used to determine the complex Fourier transform of the complex mixed return signal S.

[0141] In step 421, for the current Doppler shift of zero or more Doppler shifts observed in step 411, the FFT(S) is shifted by the Doppler shift to produce the corrected spectrum S given by equation 14a or 14b as described below. FFT As shown in Equation 27 of Foucras 2014, Doppler code compensation can be achieved by applying the time-shift theorem. In fact, the time-shift frequency theorem is given by Equation 13.

[0142]

[0143] in Let F(ζ) denote the Fourier operator, x(t) be a function of time t, δ be the time shift, and F(ζ) denote the Fourier transform of x(t). Then, for FFT-based acquisition methods, code delay caused by code Doppler can be compensated by multiplying the FFT of the locally extended code by a complex exponent in the frequency domain. The advantage of this method is that if the Fourier transform of the extended code sequence has already been constructed and stored in memory, the Fourier transform of the backward (or extended) extended code can be easily transformed to the frequency domain. The correct extended code can then be generated quickly. This technique was patented by Krasner in 1998. The effect of Doppler is to cause a spectral shift in the code. Therefore, when using the convolution theorem to quickly calculate cross-correlation, the frequency content of the measured code does not match the frequency content of the reference. Doppler compensation brings the spectrum back into alignment, and the cross-correlation is valid again.

[0144] In some implementations, Equation 14a is used to calculate the correct spectrum.

[0145] S FFT =circshift(FFT(S), ω D (14a)

[0146] Here, circshift(x, y) shifts the function x of the independent variable over a finite field by shifting the quantity y of the independent variable, such that any object shifted from one end of the finite field is shifted to the opposite end. In some implementations, the correct spectrum is calculated using Equation 14b, which removes the Doppler effect by multiplying by a complex exponent and then calculating an FFT, as shown in Equation 13.

[0147] S FFT =FFT(S*exp(-iω) D t)) (14b)

[0148] In some implementations, step 421 includes...

[0149] In step 423, the cross-correlation XC between the phase code exp(iB(t)) and the corrected complex signal Scorr is determined, designated as XC(Code, Scorr) for each of the M independent blocks with N symbols, and then averaged. In some implementations, this is done by employing the corrected complex spectrum S FFT The inverse fast Fourier transform (invFFT) is performed and the corrected complex number is returned to Scorr and correlated with the digital signal exp(iB(t)) representing the code, as given in Equation 15a.

[0150]

[0151] Where correl(x, y) is a function that determines the correlation between series x and series y, and B m (t) is the code for the m-th block. Both the invFFT and correl functions involve multiple operations on each member of the series. In some implementations, this is achieved by using S, which has already been determined in step 421. FFT Perform multiplication in Fourier space and then use inverse FFT to save computational resources, as given in Equation 15b.

[0152]

[0153] Any peak in XC(Code, Scorr) is used to determine the delay time Δt under the current Doppler offset, and zero or more delay times are used to calculate zero or more corresponding distances under the current Doppler offset.

[0154] In some implementations, FFT-based convolutions for determining cross-correlation (XC) can also be efficiently performed using convolutions based on finite impulse response (FIR) filters, as given in Equation 15c. This may be more efficient for shorter code lengths and in some computational hardware settings (FPGAs). For each distance interval k in the cross-correlation...

[0155]

[0156] Note that the dot product (*) is implied in reference code B. m This represents a series of inner products at different shifts (k) between the signal and the correction signal S. It can be seen that the FIR method in Equation 15c implies simpler register shift operations and simpler multiplications compared to the more complex FFT method in Equation 15b. For shorter code B, the repeated shifts and multiplications of the FIR method can be computationally more efficient.

[0157] Step 423 involves finding any peaks (zero or more) within the range intervals to indicate distance, if any, from which some external objects scattered the transmitted beam. The returns from each range interval will exhibit variability due to noise and speckle. The peak detection process determines whether this variability from one range interval to the next represents an actual scatterer. This determination can be difficult.

[0158] In some implementations, the phase returned from any distance interval can be correlated with the Doppler shift already detected in step 411. If the measured phase in that interval does not match the desired phase, given the known Doppler shift, the return is eliminated as the distance of the actual scatterer, and the distance interval is ignored or discarded. This is described in more detail in Example Section 4.2.

[0159] In some implementations that do not use in-phase and quadrature separation of electrical signals, the spectrum determined in step 411 has equal peaks at both positive and negative Doppler shift values ​​(e.g., at...). Figure 1D In the middle, except for f S =f0+Δf D Outside the peak at f0-Δf D There will be a second peak of equal height at the location; and the correct sign of the Doppler shift cannot be determined from the spectrum. Although equations 15a to 15c can be used to determine distances with unsigned Doppler shifts available in some embodiments (since positive or negative Doppler shifts can be used), signed Doppler shifts are still valuable, for example, in speed sensors or for vehicle control. It is found that the phase of the returned signal in equation 16a depends on the sign of the Doppler shift, where positive and negative Doppler shifts appear on two opposite sides of the unit circle depicting the phase. If the magnitude of the Doppler shift and the distance to the target are known, for example due to strong Doppler peaks and strong peaks at one or more specific distance intervals, signed Doppler can be inferred as described in more detail in Example Section 4.3. Figure 7 This is a graph illustrating an example dependence of the phase-compensated complex value of the distance peak with respect to the sign of the Doppler shift according to an embodiment. Circles represent blue-shifted Doppler datasets, x represents red-shifted Doppler datasets, and squares along the real axis represent DC (i.e., zero-Doppler shift) datasets. As can be seen, blue-shifted (positive) Doppler shift data have a positive imaginary part; red-shifted (negative) Doppler shift data have a negative imaginary part; and zero-shift data have a zero imaginary part, resulting in a real-valued distance peak.

[0160] In some implementations, the calculation of the cross-correlation with the FFT of the code signal is modified to digitally compensate for the coherent broadening in the Doppler domain, which is caused by continuous phase and frequency fluctuations due to the evolution of the phase-compiled LiDAR pulses by orders of magnitude CPI. This compensation is described in more detail below in Example Section 4.5.

[0161] In step 425, it is determined whether another Doppler frequency shift exists, for example, if more than one Doppler frequency shift is detected in step 411. If so, control returns to step 421 to use the next Doppler frequency shift to correct the complex return spectrum FFT(S). If not, control returns to step 427. In step 427, Doppler blur (if any) is removed, for example, by the overlap processing described above. There is some possibility that a so-called “split pixel” scenario may occur during scanning. In this scenario, the beam may be clipped such that one part measures a surface at one distance and Doppler, while other parts measure different distances and Dopplers. In this scenario, an efficient processing strategy is needed to extract all relevant information. For example, cross-spectrums may sense multiple non-zero Doppler values. This will result in multiple Doppler corrections and cross-correlations. One strategy is to coherently sum the time-domain signals of the Doppler corrections before a single cross-correlation. This avoids the computational burden of multiple cross-correlations at the cost of some blurring in the range-Doppler pairings and the addition of noise components to the final range profile for each correction signal. A spatial correspondence algorithm can be used to classify ambiguities. This algorithm is designed to find the "most likely" distance-Doppler pairing based on spatial proximity to non-ambiguous (single distance-Doppler) points. Additive noise may not be a problem. This processing strategy is worth considering because some users may expect more return functionality. In some implementations, step 427 is omitted and control proceeds directly to step 431.

[0162] In step 431, for example, a scan is performed to examine new points in the scene of interest to determine if another point to be illuminated exists in the scene of interest. If it exists, control returns to step 405 and subsequent steps to illuminate the next point and process any returns. In some implementations using multi-point averaging, new points are added to the average and the oldest points are removed, or P new points are collected in the loop formed in steps 405 through 409. If no other point to be illuminated exists, the result is used, and control returns to step 433.

[0163] In step 433, the device is operated based on the Doppler effect or corrected distance. In some embodiments, this involves presenting an image on a display device indicating the Doppler-corrected position of any object at a plurality of points illuminated by the transmitted optical signal. In some embodiments, this involves conveying data to the device to identify at least one object based on a point cloud of Doppler-corrected positions at a plurality of points illuminated by the transmitted optical signal. In some embodiments, this involves presenting an image on a display device indicating the magnitude of the Doppler effect at a plurality of points illuminated by the transmitted optical signal, thereby distinguishing moving objects from stationary objects and non-existent objects. In some embodiments, this involves moving a vehicle to avoid a collision with an object, wherein the closing speed between the vehicle and the object is determined based on the magnitude of the Doppler effect at a plurality of points illuminated by the transmitted optical signal. In some embodiments, this involves identifying a vehicle or identifying an object on a collision path based on a point cloud of Doppler-corrected positions at a plurality of points illuminated by the transmitted optical signal. Filtering the point cloud data based on Doppler has the effect of identifying and removing vegetation that may move in a light breeze. Hard objects, man-made objects, or dense objects are then better revealed through the filtering process. This can be advantageous in defense and surveillance scenarios. In vehicle scenarios—Doppler can be used to segment objects (i.e., road surface compared to moving vehicles).

[0164] In some implementations with multiple Doppler shifts for a single return, step 433 includes associating each delay time with one of the Doppler shifts, assuming a particular return is based on an object or part of an object moving at a specific average velocity over the duration of a transmitted signal. For a given Doppler correction, only those range peaks associated with that Doppler correction will appear in the cross-correlation. Therefore, in the case of multiple instances, it is impossible to incorrectly pair a given distance with a Doppler. In other words, the ambiguity guarantee of this method ensures that confusion is not possible.

[0165] Figure 4B This is a flowchart illustrating an example method for enhancing velocity resolution and signal-to-noise ratio in optical phase-coded distance detection according to an embodiment. Although for illustrative purposes... Figure 4B The steps are described as a whole in a specific order, but in other embodiments, one or more steps or parts thereof are performed in a different order, or overlap in time, are performed sequentially or in parallel, or are omitted, or one or more additional steps are added, or are combined in some way to change the method. In some embodiments, some or all of the operations of method 400B may be performed by processing system 350.

[0166] Method 400b includes an operation 402b of receiving an electrical signal generated by mixing a first optical signal and a second optical signal, wherein the first optical signal is generated by modulating the optical signal, and wherein the second optical signal is received in response to transmitting the first optical signal to an object. The method includes an operation 404b of determining the Doppler frequency shift of the second optical signal. The method includes an operation 406b of generating a corrected electrical signal by adjusting the electrical signal based on the Doppler frequency shift. The method includes an operation 408b of determining the distance to the object based on the cross-correlation between the corrected electrical signal and an RF signal associated with the first optical signal.

[0167] 4. Example Implementation

[0168] Figure 8 This is a block diagram illustrating an example system 801 including at least one high-resolution Doppler LIDAR system 820 mounted on a vehicle 810 according to an embodiment. In this embodiment, the LIDAR system 820 is similar to one of LIDAR systems 200, 200'. The vehicle has a center of mass represented by star 811 and travels in the forward direction given by arrow 813. In some embodiments, the vehicle 810 includes components (such as steering or braking systems (not shown)) that operate in response to signals from a processor (such as a vehicle control module 272 of processing system 250). In some embodiments, the vehicle has an onboard processor 814, such as... Figure 13 The chipset shown. In some embodiments, the onboard processor 814 communicates wirelessly with a remote processor, such as... Figure 12 As shown in the diagram. In one embodiment, the processing system 250 of the LIDAR system is communicatively coupled to the onboard processor 814, or the LIDAR processing system 250 is used to operate the onboard processor 814 such that the vehicle control module 272 causes the processing system 250 to send one or more signals to the vehicle's steering or braking system to control the vehicle's direction and speed. High-resolution Doppler LIDAR uses a scanning beam 822, represented by a future beam 823, scanning from one side to the other through an azimuth field of view 824 and through a vertical angle illuminating points around the vehicle 810. In some embodiments, the field of view is a 360-degree azimuth. In some embodiments, the tilt angle field of view is from about +10 degrees to about -10 degrees or a subset thereof.

[0169] In some embodiments, the vehicle includes auxiliary sensors (not shown), such as GPS sensors, odometers, tachometers, temperature sensors, vacuum sensors, voltage or current sensors, and other sensors known in the art. In some embodiments, a gyroscope 330 is included to provide rotation information.

[0170] In these example implementations, the LIDAR system uses the components shown above to generate phase-coded optical transmission signals. In these implementations, the symbol time (pulse width) is 2 nanoseconds (ns, 1 ns = 10^64 nm). -9 The number of symbols per block (N) is 2048, and the number of blocks M is 5. It can be used on various targets with a range from approximately 0 to approximately 250 meters, and is illuminated with beams ranging from approximately 5 to approximately 20 mm in diameter.

[0171] In various implementations, the desired type of target recognition, spatial resolution and accuracy, and object velocity resolution and accuracy are used to select the values ​​of one or more parameters of the system described above. These parameters include one or more of the following: code length, code block length, number of code blocks used for averaging, the code itself (study engineered code), shifting between the signal and the code for better detection over long ranges, speed optimization, data acquisition rate, phase modulation depth, transmitted laser power, laser spot size, scanning method, and scanning pattern.

[0172] 4.1 Enhanced speed resolution in vehicle settings

[0173] As mentioned above, the CPI (duration of a block) limits the frequency resolution of the measurement to 1 / CPI, and ultimately limits the resolution of the Doppler frequency and corresponding velocity measurements of the LIDAR system. In practice, this can limit the sensor's ability to detect the motion of a slowly moving target (such as a pedestrian) relative to a stationary target (such as a telephone pole). The situation worsens when the slowly moving target is moving perpendicular to the sensor's line of sight, further limiting the radial component of the motion. For a CPI of 3.1 μs, the frequency resolution is 323 kHz (=1 / CPI), which corresponds to a velocity resolution of approximately 0.25 m / s using Equation 2 for a 1550 nm LIDAR system (192.5 THz carrier frequency). Note that 3.1 μs CPI is the practical operating point for the experimental implementation. Therefore, if the Doppler shift is calculated on both blocks, the duration is 6.2 μs, the frequency resolution is finer at 161 kHz, and the velocity resolution will be finer at approximately 0.12 m / s; thus, it allows for the differentiation of slower-moving objects (such as pedestrians walking close to the perpendicular line of sight of the LiDAR) from stationary objects.

[0174] For scanning beam LiDAR systems, the practical upper limit of this extension of the Doppler calculation duration can be the translation of the scanning beam across multiple speckle realizations of the same object. This will result in phase evolution, which may ultimately limit the improvement of resolution. Doppler broadening compensation, described in more detail below with reference to section 4.5, addresses a possible remedy for this limitation.

[0175] The duration of the Doppler calculation (the first duration) can be dynamically adjusted to a multi-CPI extension for velocity resolution purposes, depending on the vehicle's speed, the beam's angle of view during measurement, or other operator concerns. For example, vehicles in dense urban traffic may require better pedestrian detection capabilities with finer velocity resolution for safe navigation. Alternatively, certain areas of the field of view (such as the area in front of the vehicle) may benefit from finer velocity resolution than areas to the sides of the vehicle. In summary, this approach helps to provide information conducive to the successful and safe operation of autonomous vehicles. The choice of resource utilization will be exposed to the operator of the LiDAR system.

[0176] For example, in Figure 8 In this context, stationary objects 834, such as lampposts, are advantageously distinguished from slowly moving objects 836, such as pedestrians. If a pedestrian is moving in the direction given by velocity vector 838, there is a first velocity component 842 perpendicular to the line of sight of the LiDAR system, thus contributing no Doppler shift; and there is a second velocity component 844 pointing from the LiDAR system toward the line of sight, thus contributing to the Doppler shift. If the second component is very small, such as... Figure 8 As shown, it may be difficult to distinguish the moving object 836 from the stationary object 834. The enhanced velocity resolution described in this section is very useful in this situation.

[0177] 4.2 Coherent Filtering

[0178] As mentioned above, parasitic return can be eliminated by considering the expected phase at each distance based on the observed Doppler shift. This is related to the travel time Δt and the Doppler shift Δf. D The expected phase φ of the relevant distance interval E As given by Equation 16a.

[0179] φ E =angle{exp(i2πΔf D Δt)} (16a)

[0180] This factor can be used to emphasize the real peak and reduce noise variation by further correcting the complex cross-correlation with that phase, as given in Equation 16b, so that the cross-correlation terms are rotated with the expected phase so that they point along the positive real axis.

[0181] XCcorr=XC / exp(iφ E (16b)

[0182] This method improves detection statistics because the number of noisy intervals contributing to potential false alarms is limited by additional analysis of available phase information. This effect is as follows: Figure 6A and Figure 6B As shown. Figure 6A This is a distance distribution plot showing the actual distance peak, which, according to the embodiment, is almost indistinguishable in terms of power from the noise range at greater distances. The horizontal axis represents the distance range, and the vertical axis represents the power in dB. The actual signal is marked at approximately 60 degrees of the distance range. Plotting the complex amplitudes of the various ranges illustrates better separation between noise and signal, especially when the sample under consideration is limited to a specific phase. Figure 6B This is a graph of the real and imaginary parts of the cross-correlation in Equation 16b according to the implementation method. Returns with phases similar to the expected phase in Equation 16a point along the positive real axis and are marked by hollow circles if they are within a few degrees of zero. Other returns are simple points. This graph shows that a large number of return distance intervals can be easily eliminated, making it easier to identify the detection of the actual signal.

[0183] 4.3 Signed Doppler from Coherent Filtering

[0184] As mentioned above, if the sign of the Doppler shift is unknown, the sign of the Doppler shift can be determined based on the compensated phase and the measured amplitude of the Doppler shift, either by filtering the return based on the expected phase shift or, alternatively, by filtering the return based on the expected phase shift. The phase compensation φ for the range peak... comp It is a complex conjugate φ E The phase compensation complex value of the distance peak is in Figure 7 Draw in the middle.

[0185] Figure 7 The data were collected under different conditions at an SNR of approximately 22 dB. Each experimental dataset consists of 8000 range measurements off-target from a diffusely rotating target at an angle to the beam to sample the speckle distribution and provide a Doppler shift. In doing so, it was found that a blue-shifted Doppler signal results in a phase of the range peak primarily oriented along the positive imaginary axis, a red-shifted Doppler signal results in a phase along the negative imaginary axis, and the signal at DC results in a phase along the real axis. The final result is a consequence of the fact that both the signal and the code we correlate with are real signals; therefore, if the Doppler shift is not applied, the resulting correlation must also be real. This phase correlation allows the sign of the Doppler shift to be determined from the phase of the obtained range. For example, a positive Doppler shift is assumed for compensation. If the assumption is correct, the phase of the range peak is found along the positive imaginary axis. If the assumption is incorrect, the range peak is found along the negative imaginary axis; and the sign of the Doppler shift is determined accordingly to be negative.

[0186] This technique works very well for high SNR and Doppler shifts more than two frequency resolution intervals away from DC. For smaller SNR and Doppler shifts close to DC, the phase distribution is less uniform, so small Doppler shifts may remain unsigned using this technique. In such cases, two separate measurements of in-phase and quadrature electrical signals can be used instead.

[0187] 4.4 Digitally Compensated Laser Linewidth Issues

[0188] Laser linewidth is a critical performance parameter in the system design of coherent LiDAR systems. The time (distance) delay between the transmitted signal and the local oscillator signal, as well as the coherent processing interval, both contribute to a linewidth-dependent signal-to-noise ratio (SNR) degradation. Monte Carlo simulations illustrating this trade-off space are attributed to the expected SNR loss due to imperfect laser linewidth. Phase drift of the laser on a timescale of coherent measurement (e.g., M*N*τ+Δt) causes a time delay in the return optical signal R(t) of the local oscillator. I +iR Q This decoherence between the signals ultimately disrupts the achievable coherent processing gain and negatively impacts the signal-to-noise ratio (SNR), i.e., reduces the SNR. Figure 9A and Figure 9B This is a graph showing the reduction in SNR caused by laser linewidth under two different coherent processing intervals of 2 μs and 3 μs, respectively, without compensation, according to an embodiment. The horizontal axis represents the distance range, the vertical axis represents the SNR reduction compared to an ideal laser source with negligible linewidth, and the different traces represent the range from 10 kHz (1 kHz = 10 kHz). 3 Different sampling rates are used, starting with samples per second (S / s), where each sample represents M*N symbols for a total duration D = M*N*τ; and CPI = N*τ. Due to the increased linewidth drift during longer time intervals, SNR decreases even faster for longer CPIs (larger τ).

[0189] Before determining further processing for cross-correlation, the internally reflected signal S, as described above with reference to Equation 11b, is used here. p The evolution of the laser frequency and phase detected in (t) is used to digitally compensate for the real or complex electrical signal S(t). Reference phase code B(t) and time-domain signal S p The inner product operation between time-aligned segments or their sub-segments (such as one or more blocks) of (t) produces amplitude A(t) and phase φ(t) over the processing duration due to the time delay, as given in Equations 17a and 17b.

[0190]

[0191] φ(t)=angle{A(t)} (17b)

[0192] Where np is the number of discrete samples in the time interval from t1 to t2.

[0193] Performing this measurement over time allows tracking the evolution of the laser's phase φ(t). Knowing the distance delay (Δt) of the internal loop optics... I This allows the phase evolution φ(t) to be converted into the laser's frequency drift (Δf). L As given in Equation 18a.

[0194] Δf L =Δφ / Δt I (18a)

[0195] To have a value at each time step, the slowly evolving drift detected on several samples is upsampled to the digitizer rate of the LIDAR system. The phase / frequency evolution of the laser is slower than the baud rate but faster than N*τ, so these corrections occur on such a timescale.

[0196] One type of correction is a range-independent phase correction of the local oscillator, referred to in this paper as LO correction. In this case, the time-domain signal vector S(t) is element-wise multiplied by a complex LO correction, as given in Equation 18b.

[0197] S'(t)=S(t)*exp(-iφ(t)) (18b)

[0198] The second correction is a range-related phase correction to mitigate the effects of time delay Δt. C The SNR loss in the specific distance interval of interest is represented by the center distance interval. In this case, the time-domain signal vector S(t) is element-wise multiplied by the distance correlation correction, as given in Equation 18c.

[0199] S'(t)=S(t)*exp(-i 2xf(t-Δt C )Δt C (18c)

[0200] Where f(t-Δt) C It is derived using equation 8a and shifted in time by Δt. C The frequency evolution. Assuming that the measured time evolution is roughly determined at multiple points, the time delay introduced by Equation 18c is expected to be effective over reasonable intervals of the distance range.

[0201] These two techniques are applied to illustrate SNR recovery in measurements using a laser with a suboptimal linewidth of approximately 800 kHz. Due to linewidth expansion—an SNR improvement of approximately 5 dB—the recovery is consistent with the expected SNR loss. This is illustrated in Figures 10 to 10. Figure 10C As shown in the image. Figure 10A These are graphs illustrating example distributions of signal and noise in experimental data with various linewidth corrections applied, according to different implementations. The horizontal axis represents power in dB relative to the distance peak without linewidth correction; and the vertical axis represents the maximum number of intervals with observed power levels, normalized to any power level. Traces marked with a + sign correspond to portions of intervals with only noise. The most common power is approximately -10 dB relative to the uncorrected distance peak. Dashed traces correspond to portions of intervals with signal but without correction applied. The most common power is approximately 0 dB, as expected by definition. Traces marked with hollow circles correspond to portions of intervals with only LO distance-independent correction applied. The most common power is approximately +3 dB relative to the uncorrected range peak, with an SNR recovery of 3 dB. Traces marked with an asterisk correspond to portions of intervals with both LO distance-independent correction and distance-dependent correction applied. The most common power is approximately +8 dB relative to the uncorrected distance peak, with an SNR recovery of 8 dB. This latter asterisk trace clearly shows the best depiction of the noise distribution.

[0202] Figure 10B and Figure 10C These are graphs illustrating example distance peaks in actual returned data from various implementations, applying different linewidth corrections. The horizontal axis in each indicates the distance range; and the vertical axis in each represents the power (in dB) relative to any level. Figure 10C yes Figure 10B A magnified version is shown to better distinguish the plotted traces. The dashed traces correspond to the actual return signal without applied correction. A peak with approximately 1 dB power is observed. The traces marked with hollow circles correspond to the actual return signal with only range-independent LO correction applied. A peak with approximately 3 dB power is observed. The traces marked with asterisks correspond to the actual return signal with both range-independent LO and range-dependent corrections applied. A peak with over 6 dB power is observed. Clearly, the fully corrected signal has the best SNR, which may be particularly useful for distant targets or noisy conditions.

[0203] 4.5 Digitally Compensated Doppler Frequency Broadening

[0204] Coherent broadening in the Doppler domain manifests as energy from a target in multiple adjacent frequency intervals during Doppler shift calculations. This can lead to errors in Doppler shift peak determination, which in turn can cause errors in cross-correlation and the resulting peak distance intervals.

[0205] Coherent broadening occurs due to continuous phase and frequency fluctuations, evolving on the order of magnitude of the coherent processing interval (CPI) of the phase-coded LiDAR pulse train. These phase and frequency fluctuations can originate from many different sources, including laser phase noise, plateau and target vibrations, and speckle evolution. Speckle refers to the random amplitude and phase of the coherent return due to contributions from numerous individually phased returns from a diffuse target within the diffraction-limited divergence angle. The diffraction-limited divergence angle is the minimum possible divergence for a given collimated Gaussian beam diameter—that is, physically impossible to do better than the diffraction-limited divergence angle, θ. D =λ / πw, where λ is the laser frequency and w is the Gaussian beam parameter of the collimated beam (1 / (beam radius e)). 2 Speckle evolution occurs when the contributions of these scatterers change due to beam and / or target motion. Speckle evolution is a problem in coherent beam-scanning LiDAR, which can limit scan speed to a value on the order of a diffraction-limited divergence angle per CPI, or suffer a significant loss in range peak SNR. Digital compensation for this coherent broadening enables increased beam scanning speed for the maximum range of a given target, allowing for faster update rates in coherent beam-scanning LiDAR systems.

[0206] To explain the compensation, the following is provided as a rationale. However, the method is not affected by the completeness or accuracy of the following. In coherent LIDAR, the phase of the measured signal is interference-sensitive with respect to the distance L between the scattering target and the LIDAR system, as shown in Equation 19a.

[0207] φ=4πL / λ (19a)

[0208] Where λ is the wavelength of the optical carrier. Even micrometer-scale shifts result in phase shifts, while larger-scale shifts lead to Doppler frequency shifts. For diffuse targets, interferometric sensitivity results in a speckle distribution in the received field that includes phase changes (e.g., multiple returns). For near-diffraction-limited collimated laser beams, the scale of the speckle distribution can be parameterized by the diffraction-limited divergence angle of the laser beam. Here, "scale" means the amount of angular scan displacement required to induce new speckle formation or significant changes in phase and amplitude, or both. It is the scan rate relative to the beam size (by the diffraction-limited divergence angle θ). D Limited beam divergence angle θ B The normalized parameters for the coherent processing interval (CPI) and the coherent processing interval (CPI) are expressed, for example, in degrees per second. This relationship is shown in Equation 19b.

[0209] scale = θ B / CPI (19b)

[0210] In coherent beam scanning LiDAR, rapidly scanning beyond this angle within a single coherent processing interval (CPI, e.g., N*τ) causes phase broadening of the signal. Phase broadening results in signal reduction after cross-correlation processing.

[0211] The processing method described herein partially compensates for phase broadening by utilizing Doppler domain signal (frequency) broadening as an estimate of phase evolution broadening during the coherent processing interval.

[0212] Recall that, as described above, the core processing block begins with the digitized and conditioned time-domain data, divided into array vectors S'(t) during the repetition period (N*τ) of the transmitted code. For each time-domain vector, an FFT is applied to S'(t), transforming the signal to the frequency domain, for example in Equation 12, where the Doppler signal is identified as a peak in the power spectrum or cross spectrum XS(S). The position n of the peak in the frequency space identifies the Doppler frequency of the potential target, and the power of the peak is proportional to the returned signal. The signal and phase code B... m The cross-correlation of (t) is expressed in Equation 15c by performing an FFT on the signal (e.g., S). FFT This is achieved by multiplying the phase code, which has been cyclically shifted to align the DC frequency value with the Doppler frequency of the signal, by the conjugate of the FFT. In other implementations, other versions of the multiplication are performed, for example, as shown in Equations 15a or 15b. After multiplication, an inverse FFT is applied to calculate the cross-correlation between the time-domain signal and the frequency-shifted code, thereby producing the Doppler compensation and the complete range profile of the signal data vector. This process is repeated at each transmit / receive interval (N*τ). In some implementations, when M > 1, the results of M such calculations are averaged.

[0213] According to the illustrated implementation, the Doppler-compensated cross-correlation of the signal and phase code is repeated K times, each time by multiplying the Doppler-compensated signal with a shifted, scaled, and phase-fixed version of the FFT of the code Bm(t). These results are then coherently summed before performing the inverse FFT to compute the compensated Doppler-broadened range profile. It is important to note that this method is not equivalent to compensating for different Doppler targets. For this Doppler-broadening compensation technique to improve the SNR of target signals in the range domain, the Doppler compensation is treated collectively as a single signal from a single target. The choice of shift, scaling, and phase can cause constructive or destructive interference of the range peaks.

[0214] Figure 11A and Figure 11B This is a spectrum illustrating the example effects of frequency broadening due to speckle on the selection of the Doppler peak at two different sampling rates, according to various embodiments. The horizontal axis indicates the spectral (frequency shift) interval, and the vertical axis indicates the power in dB relative to the minimum observed power. The solid line represents the peak with... Figure 11A 14,400 degrees per second and Figure 11B Simulated data of rapid speckle evolution at a scan rate of 3,360 degrees per second. For a beam diverging by 100 microradians, this corresponds to a single 3-microsecond CPI used in many example implementations. Figure 11A Approximately 7.4 speckle units and Figure 11B The 1.8 speckle units in the diagram result in two main effects. First, the peak height of the signal is reduced, making thresholding of the Doppler signal in a shot noise background more difficult. Second, the energy spread between multiple Doppler intervals means that the cross-correlated code energy is also shifted between more than one Doppler interval, making individual Doppler compensation shifts not fully correlated, thus reducing the final SNR of the distance peak. In each figure, two or three distinct peaks are evident due to coherent broadening caused by the simulated speckle evolution. The dashed lines represent a low-pass Gaussian filtered version of the same spectrum signal. A K=5 Doppler frequency shift centered on the peak in the dashed trace (the low-pass version of the signal) is used for digital compensation and is represented by hollow circles.

[0215] Typically, to compensate for Doppler broadening, the first step is to correctly identify the Doppler peaks in the power spectrum from the shot noise background. In a preferred embodiment, a finite impulse response (FIR) filter G is used to smooth the Doppler spectrum, producing a dashed trace in each plot. This smoothing helps to integrate the energy over the speckle broadening peaks so that they are higher than the fluctuations in the shot noise background. In a preferred embodiment, the filter is defined by a radius r, which sets the number of taps in the filter to 2r+1 and sets the half-width of the filter itself as 1 / e of a Gaussian shape to √2r. After applying the filter, assuming it is in the Doppler region of interest above a threshold power level (e.g., average background noise), the Doppler peaks are clearly the interval n with the maximum power in the dashed trace. Then, the K maximum points in the original signal (solid trace) within a radius r around interval n (specified as n) are identified. k The interval (in which the interval is selected) is used for Doppler compensation.

[0216] Then, the Doppler broadened compensated cross-correlation distance profile is calculated according to Equation 20.

[0217]

[0218] The fraction is a scaling factor, and the exponent is a phase factor that takes into account the accumulated phase across different Doppler frequency intervals of interval Δf, the reference round-trip time to the target, and the corresponding reference distance L. r The return Δt rThis approach is similar to deconvolution in the frequency domain before a matched filter. The range-dependent phase factor means that the effectiveness of the compensation algorithm depends on the choice of the surrounding distances to be compensated and the actual distances. If the distance to be corrected is poorly chosen, this will result in a deterioration of the SNR at the range peak rather than an increase in the final SNR. In LiDAR, the SNR at shorter distances is generally better than the SNR at longer distances, thus suggesting that if a single distance is chosen for compensation, it should be close to the maximum expected distance, as shorter distances typically have sufficient SNR to allow for some loss at these wavelengths without significantly reducing the detection probability.

[0219] Figure 11C and Figure 11D These are range plots illustrating example effects of digital compensation for frequency broadening at two different sampling rates, according to various embodiments. The horizontal axis represents the range interval (L) in meters, and the vertical axis represents power in arbitrary units on a linear scale. The original range profile is indicated by a solid trace; and the range profile after compensation for coherent broadening caused by simulated speckle is represented by a dashed line in each plot, where... Figure 11C Corresponding to Figure 11A The simulated data in Figure 11D Corresponding to Figure 11B The simulation data is from [source name]. For each scan rate, the compensated distance profile shows a significantly stronger peak at 200m. For [data missing] Figure 11C The higher scan rate allows for compensation to provide the difference between missing and detected targets at 200m.

[0220] Figure 11E and Figure 11F These are graphs illustrating example improvements in signal-to-noise ratio (SNR) as a function of distance, resulting from digital compensation for frequency broadening with two different choices of the number of Doppler peaks used, according to various embodiments. In each graph, the horizontal axis indicates distance in meters; while the vertical axis represents the average SNR in dB. These graphs show the results of applying speckle compensation to simulated speckle broadening data relative to distance for a given chosen compensation radius r and different numbers K of frequency points to be compensated. Each trace indicates the SNR when the simulated target moves from 0 to 275 meters. Traces with closed circles represent the result without coherent broadening compensation. Traces marked with closed squares represent the result with compensated coherent broadening, which is for a distance L corresponding to 125 meters to the target. r And tuned. The trace with a closed triangle indicates the use of compensated gain. At the reference distance L r At 125m, the gain is maximum, approximately 2dB. There is a loss within 50 meters, but no gain beyond 200 meters. Compare this to K=3. Figure 11E and K=5 Figure 11FThese figures show that the maximum gain of the speckle correction algorithm increases with the number K of Doppler frequency compensations. However, when the actual distance peak is far from the assumed reference distance L in the compensation... r In this context, the increased gain also means the increased loss.

[0221] 5. Overview of Computing Hardware

[0222] Figure 12 This is a block diagram illustrating a computer system 1200 on which embodiments of the present disclosure may be implemented. The computer system 1200 includes communication mechanisms, such as a bus 1210 for transmitting information between other internal and external components of the computer system 1200. Information is represented as a physical signal of a measurable phenomenon, typically voltage, but in other embodiments includes phenomena such as magnetic, electromagnetic, pressure, chemical, molecular, atomic, and quantum interactions. For example, north and south magnetic fields or zero and non-zero voltages represent two states (0, 1) of a binary digit (bit). Other phenomena may represent higher cardinality numbers. A superposition of multiple simultaneous quantum states prior to measurement represents a qubit (quantum bit). A sequence of one or more digits constitutes digital data for representing numbers or codes. In some embodiments, information called analog data is represented by an approximately continuous region of measurable values ​​within a specific range. The computer system 1200 or a portion thereof constitutes means for performing one or more steps of one or more methods described herein.

[0223] A sequence of binary bits constitutes digital data used to represent numbers or codes of characters. Bus 1210 includes a plurality of parallel information conductors, enabling rapid transfer of information between devices coupled to bus 1210. One or more processors 1202 are coupled to bus 1210 for processing information. Processor 1202 performs a set of operations on the information. This set of operations includes introducing information from bus 1210 and placing information on bus 1210. This set of operations typically also includes comparing two or more information units, shifting the positions of information units, and combining two or more information units, such as by addition or multiplication. A series of operations to be executed by processor 1202 constitutes computer instructions.

[0224] Computer system 1200 also includes memory 1204 coupled to bus 1210. Memory 1204, such as random access memory (RAM) or other dynamic storage devices, stores information including computer instructions. Dynamic memory allows computer system 1200 to change the information stored therein. RAM allows information units stored at locations called memory addresses to be stored and retrieved independently of information at adjacent addresses. Memory 1204 is also used by processor 1202 to store temporary values ​​during the execution of computer instructions. Computer system 1200 also includes read-only memory (ROM) 1206 or other static storage devices coupled to bus 1210 for storing static information, including instructions, that is not changed by computer system 1200. Also coupled to bus 1210 is a non-volatile (permanent) storage device 1208, such as a magnetic disk or optical disk, for storing information, including instructions, that remains even when computer system 1200 is turned off or otherwise powered down.

[0225] Information, including instructions, is provided from external input devices 1212 to the bus 1210 for use by the processor. These external input devices include, for example, a keyboard with alphanumeric keys operated by a human user, or sensors. Sensors detect nearby conditions and convert those detections into signals compatible with signals used to represent information within the computer system 1200. Other primary external devices coupled to the bus 1210 for human interaction include display devices 1214 for presenting images, such as cathode ray tubes (CRTs) or liquid crystal displays (LCDs), and pointing devices 1216, such as mice, trackballs, or cursor arrow keys, for controlling the position of a small cursor image presented on the display 1214 and issuing commands associated with graphical elements presented on the display 1214.

[0226] In the illustrated embodiment, dedicated hardware, such as application-specific integrated circuit (IC) 1220, is coupled to bus 1210. The dedicated hardware is configured to perform operations that the processor 1202 cannot perform at a sufficiently fast speed for a specific purpose. Examples of dedicated ICs include graphics accelerator cards for generating images for display 1214, cryptographic boards for encrypting and decrypting messages sent over a network, speech recognition, and interfaces for special external devices, such as robotic arms and medical scanning equipment that repeatedly perform complex sequences of operations that are more efficiently implemented in hardware.

[0227] Computer system 1200 also includes one or more instances of a communication interface 1270 coupled to bus 1210. Communication interface 1270 provides bidirectional communication to various external devices that operate using their own processors, such as printers, scanners, and external disks. Generally, coupling utilizes network link 1278, which connects to local network 1280, to which various external devices with their own processors are connected. For example, communication interface 1270 may be a parallel port, a serial port, or a Universal Serial Bus (USB) port on a personal computer. In some embodiments, communication interface 1270 is an Integrated Services Digital Network (ISDN) card, a Digital Subscriber Line (DSL) card, or a telephone modem that provides information communication connectivity to a corresponding type of telephone line. In some embodiments, communication interface 1270 is a cable modem that converts signals on bus 1210 into signals for communication connections over coaxial cable or into optical signals for communication connections over fiber optic cable. As another example, communication interface 1270 may be a Local Area Network (LAN) card that provides data communication connectivity to a compatible LAN, such as Ethernet. Wireless links may also be implemented. Carrier waves, such as sound waves and electromagnetic waves, include radio, optical, and infrared waves, which travel through space without wires or cables. Signals include artificial variations in the amplitude, frequency, phase, polarization, or other physical properties of the carrier wave. For wireless links, communication interface 1270 transmits and receives electrical, acoustic, or electromagnetic signals, including infrared and optical signals, carrying streams of information such as digital data.

[0228] The term "computer-readable medium" is used herein to refer to any medium that participates in providing information to processor 1202, including instructions for execution. Such media can take many forms, including but not limited to non-volatile media, volatile media, and transmission media. Non-volatile media include, for example, optical discs or magnetic disks, such as storage device 1208. Volatile media include, for example, dynamic memory 1204. Transmission media include, for example, coaxial cables, copper wires, fiber optic cables, and waves that travel through space without wires or cables, such as sound waves and electromagnetic waves, including radio waves, optical waves, and infrared waves. The term "computer-readable storage medium" is used herein to refer to any medium other than transmission media that participates in providing information to processor 1202.

[0229] Common forms of computer-readable media include, for example, floppy disks, flexible disks, hard disks, magnetic tape or any other magnetic media, optical disc ROM (CD-ROM), digital video disc (DVD) or any other optical media, punched cards, paper tape or any other physical media with a perforated pattern, RAM, programmable ROM (PROM), erasable ROM (EPROM), FLASH-EPROM or any other memory chip or cassette tape, carrier waves or any other media that a computer can read. The term "non-transitory computer-readable storage medium" is used herein to refer to any medium, other than carrier waves and other signals, that participates in providing information to processor 1202.

[0230] The logic encoded in one or more tangible media includes one or both of computer-readable storage media and processor instructions on dedicated hardware such as the ASIC 1220.

[0231] Network link 1278 typically provides information communication to other devices using or processing information via one or more networks. For example, network link 1278 may provide a connection via local network 1280 to host computer 1282 or equipment 1284 operated by an Internet service provider (ISP). ISP equipment 1284, in turn, provides data communication services via a public global packet-switched communications network, now commonly referred to as the Internet 1290. A computer connected to the Internet, referred to as server 1292, provides services in response to information received via the Internet. For example, server 1292 provides information representing video data for presentation at display 1214.

[0232] This disclosure relates to the use of a computer system 1200 for implementing the techniques described herein. According to one embodiment of this disclosure, those techniques are performed by the computer system 1200 in response to a processor 1202 executing one or more sequences of one or more instructions contained in memory 1204. Such instructions (also referred to as software and program code) may be read into memory 1204 from another computer-readable medium, such as storage device 1208. Execution of the sequence of instructions contained in memory 1204 causes the processor 1202 to perform the method steps described herein. In alternative embodiments, hardware such as application-specific integrated circuit 1220 may be used instead of or in combination with software to implement this disclosure. Therefore, embodiments of this disclosure are not limited to any particular combination of hardware and software.

[0233] Information carried to and from computer system 1200 is transmitted via network link 1278 and other networks through communication interface 1270. Computer system 1200 can send and receive information including program code via network link 1278 and communication interface 1270, via networks 1280, 1290, etc. In an example using Internet 1290, server 1292 sends program code for a specific application requested by a message sent by computer 1200 via Internet 1290, ISP equipment 1284, local network 1280, and communication interface 1270. The received code can be executed by processor 1202 upon receipt, or it can be stored in storage device 1208 or other non-volatile memory, or both, for later execution. In this way, computer system 1200 can obtain application code in the form of signals on a carrier wave.

[0234] Carrying one or more sequences of instructions or data, or both, to processor 1202 for execution may involve various forms of computer-readable media. For example, instructions and data may initially be carried on a disk of a remote computer, such as host 1282. The remote computer loads the instructions and data into its dynamic memory and transmits the instructions and data over a telephone line using a modem. A modem local to computer system 1200 receives the instructions and data over the telephone line and uses an infrared transmitter to convert the instructions and data into signals on an infrared carrier wave used as network link 1278. An infrared detector, used as communication interface 1270, receives the instructions and data carried in the infrared signal and places information representing the instructions and data on bus 1210. Bus 1210 carries the information to memory 1204, from which processor 1202 retrieves and executes the instructions using some data sent along with the instructions. The instructions and data received in memory 1204 may optionally be stored on storage device 1208 before or after execution by processor 1202.

[0235] Figure 13 A chipset 1300 is shown on which embodiments of the present disclosure may be implemented. The chipset 1300 is programmed to perform one or more steps of the methods described herein, and includes, for example, information incorporated in one or more physical packages (e.g., chips). Figure 13 The processor and memory components described herein. By way of example, a physical package includes an arrangement of one or more materials, components, and / or lines on a structural component (e.g., a substrate) to provide one or more properties, such as physical strength, size retention, and / or electrical interaction limitation. It is contemplated in some embodiments that the chipset may be implemented in a single chip. Chipset 1300 or a portion thereof constitutes means for performing one or more steps of the methods described herein.

[0236] In one implementation, chipset 1300 includes communication mechanisms, such as a bus 1301 for transferring information between components of chipset 1300. Processor 1303 has a connection to bus 1301 to execute instructions and process information stored, for example, in memory 1305. Processor 1303 may include one or more processing cores, each configured to operate independently. Multi-core processors are capable of multiprocessing within a single physical package. Examples of multi-core processors include two, four, eight, or more processing cores. Alternatively or additionally, processor 1303 may include one or more microprocessors arranged sequentially via bus 1301 to enable independent instruction execution, pipelined processing, and multithreading. Processor 1303 may also be accompanied by one or more dedicated components, such as one or more digital signal processors (DSPs) 1307 or one or more application-specific integrated circuits (ASICs) 1309, to perform certain processing functions and tasks. DSP 1307 is typically configured to process real-world signals (e.g., sound) in real time independently of processor 1303. Similarly, the ASIC 109 can be configured to perform specialized functions that are not easily performed by general-purpose processors. Other specialized components that help perform the functions of the invention described herein include one or more field-programmable gate arrays (FPGAs) (not shown), one or more controllers (not shown), or one or more other specialized computer chips.

[0237] Processor 1303 and its associated components have a connection to memory 1305 via bus 1301. Memory 1305 includes dynamic memory (e.g., RAM, disk, writable optical disc, etc.) and static memory (e.g., ROM, CD-ROM, etc.) for storing executable instructions that, when executed, perform one or more steps of the methods described herein. Memory 1305 also stores data associated with or generated by performing one or more steps of the methods described herein.

[0238] 6. Modify, expand, and alter

[0239] This disclosure has been described with reference to specific embodiments thereof. However, it will be apparent that various modifications and changes may be made therein without departing from the broader spirit and scope of this disclosure. Therefore, the specification and drawings should be considered illustrative rather than restrictive. Throughout the specification and claims, unless the context otherwise requires, the words “comprising” and variations thereof, such as “including” and “having”, will be understood to imply the inclusion of an item, element, or operation, or a group of items, elements, or operations, but do not exclude any other item, element, or operation or group of items, elements, or operations. Furthermore, the indefinite article “a / an” is intended to indicate one or more items, elements, or operations modified by the article.

[0240] Although the numerical ranges and parameters used to illustrate wide ranges are approximations, the numerical values ​​illustrated in the specific, non-limiting examples are reported as precisely as possible. However, any numerical value inherently contains some error, which is necessarily caused by the standard deviation found in their respective test measurements at the time of writing. Furthermore, unless clearly apparent from the context, the numerical values ​​presented herein have an implicit precision given by the least significant digit. Thus, the value 1.1 means a value from 1.05 to 1.15. The term “about” is used to indicate a wider range centered on a given value, meaning a wider range near the least significant digit unless clearly apparent from the context; for example, “about 1.1” means a range from 1.0 to 1.2. If the least significant digit is unclear, the term “about” means twice; for example, “about X” means a value in the range from 0.5X to 2X, and “about 100” means a value in the range from 50 to 200. Furthermore, all ranges disclosed herein should be understood to include any and all subranges contained herein. For example, the range of “less than 10” can include any and all subranges between (and including) the minimum value of zero and the maximum value of 10, that is, any and all subranges with a minimum value equal to or greater than zero and a maximum value equal to or less than 10, such as 1 to 4.

[0241] Some embodiments of this disclosure are described below in the context of binary π / 2 (90-degree) phase encoding at the radio frequency modulated onto an optical signal; however, embodiments are not limited to this context. For example, in other embodiments, other phase encodings with different phase differences (e.g., 30, 60, or 180 degrees) or encodings with three or more different phases are used. Embodiments have been described in the context of a single beam and its return to a single detector or detector pair; in other embodiments, any known scanning device, such as linearly stepping or rotating optical components, or utilizing a transmitter array or detector array or detector pair, can be used to scan a single detector or detector pair. For purposes of description, "phase code duration" refers to the duration of a code indicating the phase sequence of a phase-encoded signal modulated onto an optical signal.

Claims

1. An autonomous vehicle control system, comprising: one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to: receive an electrical signal generated based on a return optical signal reflected from an object; transform the electrical signal into a frequency domain to generate a transformed electrical signal; determine a Doppler shift of the return optical signal based on a cross-spectrum of the transformed electrical signal; determine whether the object is moving closer or further away from the autonomous vehicle control system based on the Doppler shift, wherein the object is moving closer to the autonomous vehicle control system corresponds to a blue shift in the cross-spectrum, the blue shift being a positive Doppler shift, and wherein the object is moving further away from the autonomous vehicle control system corresponds to a red shift in the cross-spectrum, the red shift being a negative Doppler shift; based on the signed Doppler shift, compute a cross-correlation by multiplying the transformed electrical signal with a shifted version of a Fourier transform of a phase code, a scaled version of the Fourier transform of the phase code, and a phased version of the Fourier transform of the phase code; and determine a distance to the object based on the cross-correlation, and control at least one of a steering system and a braking system based on the determined distance.

2. The autonomous vehicle control system of claim 1, wherein the one or more processors are further configured to: determine the Doppler shift of the return optical signal based on a particular point in the cross-spectrum of the transformed electrical signal.

3. The autonomous vehicle control system of claim 1, wherein the one or more processors are further configured to: generate the cross-spectrum of the transformed electrical signal by processing the transformed electrical signal and a conjugate of the Fourier transform of the phase code.

4. The autonomous vehicle control system of claim 3, wherein the one or more processors are further configured to: circularly shift the phase code to align a direct current (DC) frequency with a Doppler frequency associated with the electrical signal.

5. The autonomous vehicle control system of claim 1, wherein the one or more processors are further configured to: recompute the cross-correlation according to an interval associated with transmission or reception of an optical signal.

6. The autonomous vehicle control system of claim 1, wherein the one or more processors are further configured to: adjust the electrical signal by removing the Doppler shift from the electrical signal to generate an adjusted electrical signal; and determine a distance to the object based on the adjusted electrical signal.

7. The autonomous vehicle control system of claim 1, wherein the one or more processors are further configured to: determine an in-phase component and a quadrature component of the return optical signal based on the electrical signal; and determine a separation of the in-phase component and the quadrature component.

8. The autonomous vehicle control system of claim 7, wherein the one or more processors are further configured to: determine a sign of the Doppler shift based on the separation.

9. A light detection and ranging (LIDAR) system, the LIDAR system comprising: one or more processors; and one or more computer-readable storage media storing instructions that, when executed by the one or more processors, cause the one or more processors to: receive an electrical signal generated based on a return optical signal reflected from an object; transform the electrical signal into a frequency domain to generate a transformed electrical signal; determine a Doppler shift of the return optical signal based on a cross-spectrum of the transformed electrical signal; determine whether the object is moving closer or further away from an autonomous vehicle control system based on the Doppler shift, wherein the object is moving closer to the autonomous vehicle control system corresponds to a blue shift in the cross-spectrum, the blue shift being a positive Doppler shift, and wherein the object is moving further away from the autonomous vehicle control system corresponds to a red shift in the cross-spectrum, the red shift being a negative Doppler shift; and based on the signed Doppler shift, compute a cross-correlation by multiplying the transformed electrical signal with a shifted version of a Fourier transform of a phase code, a scaled version of the Fourier transform of the phase code, and a phased version of the Fourier transform of the phase code.

10. An autonomous vehicle, comprising: at least one of a steering system and a braking system; and a vehicle controller, the vehicle controller comprising one or more processors configured to: receive an electrical signal generated based on a return optical signal reflected from an object; transform the electrical signal into a frequency domain to generate a transformed electrical signal; determine a Doppler shift of the return optical signal based on a cross-spectrum of the transformed electrical signal; determine whether the object is moving closer or further away from the autonomous vehicle based on the Doppler shift, wherein the object is moving closer to the autonomous vehicle control system corresponds to a blue shift in the cross-spectrum, the blue shift being a positive Doppler shift, and wherein the object is moving further away from the autonomous vehicle control system corresponds to a red shift in the cross-spectrum, the red shift being a negative Doppler shift; based on the signed Doppler shift, compute a cross-correlation by multiplying the transformed electrical signal with a shifted version of a Fourier transform of a phase code, a scaled version of the Fourier transform of the phase code, and a phased version of the Fourier transform of the phase code; and determine a distance to the object based on the cross-correlation, and control the at least one of the steering system and the braking system based on the determined distance.

11. The autonomous vehicle of claim 10, wherein the one or more processors are further configured to: determine the Doppler shift of the return optical signal based on a particular point in the cross-spectrum of the transformed electrical signal.

12. The autonomous vehicle of claim 10, wherein the one or more processors are further configured to: generate the cross-spectrum of the transformed electrical signal by processing the transformed electrical signal and a conjugate of the Fourier transform of the phase code.

13. The autonomous vehicle of claim 12, wherein the one or more processors are further configured to: cyclically shift the phase code to align a direct current (DC) frequency with a Doppler frequency associated with the electrical signal.

14. The autonomous vehicle of claim 10, wherein the one or more processors are further configured to: recompute the cross-correlation according to an interval associated with transmission or reception of the optical signal.

15. The autonomous vehicle of claim 10, wherein the one or more processors are further configured to: adjust the electrical signal to generate an adjusted electrical signal by removing the Doppler shift from the electrical signal; and determine a distance to the object based on the adjusted electrical signal.

16. The autonomous vehicle of claim 10, wherein the one or more processors are further configured to: determine an in-phase component and a quadrature component of the return optical signal based on the electrical signal; determine a separation of the in-phase component and the quadrature component; and determine a sign of the Doppler shift based on the separation.

Images