Aeva, Inc.

Jose Krause Perin

Kumar Bhargav Viswanatha

Rajendra Tushar Moorti

Mina A Rezk

Esha John

Krishna Toshniwal

James Nakamura

Bruno Hexsel

Carlo Giustini

Amol V Gole

Youye Xie

Khachatur Armenyan

James Reuther

Da An

Kshitij Jain

A light detection and ranging (LIDAR) system performs a method including generating a frequency domain waveform based on a baseband signal in a time domain, determining a first likelihood metric for frequencies in the frequency domain waveform, and identifying one or more frequencies in the frequency domain waveform that exceed a threshold value for the first likelihood metric. The method further includes determining a second likelihood metric for the frequencies in the frequency domain waveform, selecting a peak frequency from the frequency domain waveform corresponding to the frequency with the highest value for the second likelihood metric based on the one or more frequencies in the frequency domain waveform that exceed the threshold value for the first likelihood metric, and determining one or more properties of a target based at least in part on the selected peak frequency and the corresponding values of the first and second likelihood metrics.

Techniques for peak detection in a LIDAR system

A set of POIs of a point cloud are received at a first filter. Each POI of the set of POIs is filtered. At a second filter, based on a first metric, a first score of the POI is determined. At a third filter, based on a second metric, a second score of the POI is determined. At the first filter, based on the first score and the second score, whether to accept the POI, modify the POI, or reject the POI, is determined to extract range or velocity information.

Techniques for parallel cascade filtering in point clouds

A set of POIs of a point cloud are received at a first filter, where each POI of the set of POIs comprises one or more points. Each POI of the set of POIs is filtered. A set of neighborhood points of a POI is selected. A metric for the set of neighborhood points is computed based on a property of the set of neighborhood points and the POI, wherein the property comprises a velocity. Based on the metric, whether to accept the POI, modify the POI, reject the POI, or transmit the POI to a second filter, to extract at least one of range or velocity information related to the target is determined. Provided the POI is not accepted, modified, or rejected, the POI is transmitted to the second filter to determine whether to accept, modify, or reject the POI to extract the at least one of range or velocity information related to the target.

Techniques for filtering points in a light detection and ranging system

A method to compensate for phase impairments in a light detection and ranging (LIDAR) system includes estimating one or more phase impairments in the LIDAR system using a digitally-sampled reference signal to produce one or more estimated phase impairments and performing one or more corrections on one or more phase impairments in a digitally-sampled target signal based on the one or more estimated phase impairments.

Compensating for phase impairments in LIDAR systems

A method of compensation in a light detection and ranging (LIDAR) system. The method includes applying a first frequency shift to a target signal to compensate for doppler shift in the target signal and performing a phase impairment correction on the target signal to produce a corrected target signal. The method further includes undoing the first frequency shift on the corrected target signal.

Techniques for correcting phase impairments in a target signal

A method of determining range and velocity information of a target in a light detection and ranging (LiDAR) system, includes determining a set of signal power density spectrums in a frequency domain based on an optical signal returned from a target, wherein a number of the set of signal power density spectrums is determined based on an expected mirror Doppler shift associated with the optical signal returned from the target. The method further includes averaging the set of signal power density spectrums to generate a compensated signal power density spectrum and detecting a peak value in the compensated signal power density spectrum to determine range and velocity information related to a target based on a corresponding frequency of the peak value of compensated signal power density spectrum.

Techniques to use power spectrum density in coherent LiDAR systems

A light detection and ranging (LIDAR) system includes a processor and a memory. The memory stores a plurality of data points and stores instructions that cause the LIDAR system to: generate the plurality of data points associated with one or more return beams corresponding to one or more optical beams transmitted towards a target; perform a plurality of processing operations on the plurality of data points to generate a point cloud corresponding to the target, wherein a first processing operation of the plurality of processing operations is configured to output a pair of indices as input to a second processing operation of the plurality of processing operations, the pair of indices referring to memory locations of a first data point and a second data point of the plurality of data points, respectively; and calculate a range and a velocity of the target based on the point cloud.

Techniques for point cloud processing utilizing point indices

A light detection and ranging (LIDAR) system transmits, towards a target, a set of chirp signals. The LIDAR system receives from the target, a set of adjusted chirp signals. The LIDAR system then determines, based on the set of adjusted chirp signals, a degree of ghosting mitigation to compensate for a ghost target appearing in a point cloud at a location where no real target exists.

Techniques for ghosting mitigation in coherent LIDAR systems using multiple chirp rates

A frequency modulated continuous wave (FMCW) light detection and ranging (LIDAR) system includes a processor and a memory. The memory stores instructions that, when executed by the processor, cause the system to: generate subbands in a frequency domain based on a range-dependent time domain baseband signal, classify each subband into a subband type, select processing parameters for each subband based on the respective subband type, and process each of the subbands using the selected processing parameters for the subband.

Signal processing based on signal subband typing

A method includes transmitting a plurality of optical beams towards a plurality of targets, receiving a plurality of return signals based on reflections of the plurality of optical beams from the plurality of targets, and generating a first plurality of peaks each associated with a different up-chirp frequency of the plurality of optical beams and a second plurality of peaks each associated with a different down-chirp frequency of the plurality of optical beams. The method further includes determining peak shape similarities between each of the first plurality of peaks and the second plurality of peaks, pairing each peak of the first plurality of peaks with a peak of the second plurality of peaks based on the peak shape similarities, and identifying the plurality of targets based on the pairing.

Techniques to associate peaks in multi-target scenarios in coherent LIDAR systems

A light detection and ranging (LIDAR) system including a processor to receive a return signal from a target based on an optical beam transmitted towards the target and receive a baseband signal in a time domain based on the return signal. The processor of the LIDAR system further to produce a comparison of signal peaks of the baseband signal with an estimate of LIDAR system noise in the frequency domain, and identify targets based on the comparison.

LIDAR system target detection

A return signal from a target is received based on an optical beam from an optical source of a LiDAR system. The return signal is sampled and converted to a frequency domain, where the return signal comprises a first frequency waveform. A matched filter is selected, where the matched filter comprises a second frequency waveform to match the first frequency waveform. The matched filter is updated by updating a set of coefficients of the second frequency waveform. The return signal is filtered by the updated matched filter to generate a filtered return signal to extract range and velocity information of the target.

Techniques for using matched filtering in coherent LiDAR systems

A light detection and ranging (LIDAR) system to generate a baseband signal in the time domain from the return signal, where the baseband signal includes frequencies corresponding to LIDAR target ranges; and a signal processing system coupled with the optical processing system to measure energy of the baseband signal in the frequency domain, to compare the energy to an estimate of LIDAR system noise, and to determine a likelihood that an signal peak in the frequency domain indicates a detected target.

Techniques for LIDAR system noise calibration and compensation for improved target detection

A light detection and ranging (LIDAR) system to transmit optical beams including at least up-chirp frequency and at least one down-chirp frequency toward targets in a field of view of the LIDAR system and receive returned signals of the up-chirp and the down-chirp as reflected from the targets. The LIDAR system may determine multiple frequency peaks associated with the target based on the returned signals. Upon determining that at least one of the multiple frequency peaks is within one or more sets of frequency ranges, the LIDAR system may combine an in-phase signal and a quadrature signal of the returned signals to generate a complex signal that enables determining whether the at least one of the multiple frequency peaks is associated with ghosting. Upon determining to be free from ghosting, the LIDAR system determines one or more of the target location, a target velocity, and a target reflectivity.

Techniques for ghosting mitigation in coherent LIDAR systems using in-phase/quadrature phase (IQ) processing

A number of measurements of an input spectrum is determined based on a scan mirror speed of the LiDAR system and a predetermined accuracy threshold in the number of measurements of the input spectrum. A set of signals are sampled at the LiDAR system and the set of signals are converted to a frequency domain to generate a set of sampled signals in the frequency domain. The set of signals are received consecutively over time. A set of first functions are created based on the set of sampled signals. The set of first functions are averaged to generate a second function. The second function represents a power spectrum density estimate of the set of signals. A peak value of the second function is detected to determine range and velocity information related to a target based on a corresponding frequency of the peak value of the second function.

Growing community of inventors

Mountain View, CA, United States of America

Jose Krause Perin