Abstract: Various embodiments relate to a system and method of pruning a machine learning model, including: training the machine learning model using training input data; calculating alpha values for different parts of the machine learning model based on gradients used in training the machine learning model wherein the alpha values are an importance metric; accumulating the calculated alpha values across training iterations; and pruning the machine learning model based upon the accumulated alpha values.

Abstract: An automotive radar system and method are configured to transmit and receive radar signals. A first received radar signal is processed to generate a range-Doppler data frame. A first target cluster is identified at a first range in the range-Doppler data frame. A range spectrum data set associated with the first range is extracted from the range-Doppler data frame. A low-pass filter is applied to the range spectrum data set to extract a first portion of a spectrum of the range spectrum data set and an inverse fast Fourier transform (IFFT) of the first portion of the spectrum is performed to generate a time-domain set of signal magnitudes. A super-resolution spectral estimation is applied to the time-domain set of signal magnitudes to identify a first range of a first target associated with the first target cluster. The first range is transmitted to a vehicle controller.

Abstract: Aspects of the present disclosure are directed to radar communications with disparate pulse repetition intervals, as may be implemented with radar transmission, receiver and processing circuitry. As may be utilized in accordance with one or more embodiments herein, time division multiplexing (TDM) multi-input multi-output (MIMO) radar signals are transmitted by transmitting sets of successive radar signals, each set having a pulse repetition interval (PRI) that is different than the PRI of sets of radar signals transmitted in another one of the sets. Positional characteristics of a target may be ascertained based on the PRI used in each of the sets and on phase characteristics of ones of the radar signals reflected from the target.

Abstract: A vehicle radar system, apparatus and method use a radar control processing unit to generate a target response signal in at least a first dimension from compressed radar data signals and to perform cell-averaging constant false alarm rate (CA-CFAR) target detection by convolving the target response signal with a weighted kernel window signal in a frequency domain using a Fast Fourier Transform hardware accelerator, an element-wise multiplier, and an Inverse Fast Fourier Transform hardware accelerator to generate an output signal having a sign that indicates a target detection decision.

Abstract: A device for a radar sensor is disclosed, the device comprising: transmission circuitry configured to generate transmission signals with a linear frequency chirp modulation in a predetermined frequency band for output to a radar antenna; reception circuitry configured to receive reflection signals corresponding to reflection of the transmitted radar signals from one or more physical objects; and control circuitry configured to select a frequency range within said predetermined frequency band and/or a timing pattern for said transmission signals; wherein said device is configured to: receive a further signal from a further radar sensor; determine, from said further signal, a frequency range and/or timing pattern in use by said further radar sensor for transmission of further transmission signals; and select a frequency range within said predetermined frequency band and/or a timing pattern for said transmission signals which does not conflict with the frequency range and/or timing pattern of said further transm

Abstract: A radar system includes transmitter and receiver antennas positioned to illuminate and receive reflections from a ground surface, a processor, and a non-transitory computer-readable medium. The processor obtains ground reflections, the corresponding ranges, and measured radial velocities, and determines a set of test ego velocities. For each test ego velocity and ground reflection, the processor generates a test radial velocity. The processor determines an absolute difference between the test radial velocity and the measured radial velocity, and whether the absolute difference satisfies a criterion. In response to satisfying the criterion, the absolute difference is accumulated into a total cost for the test ego velocity. After each ground reflection and test ego velocity is analyzed, the processor compares the total costs for the test ego velocities to obtain a smallest total cost and corresponding test ego velocity. The test ego velocity is an ego velocity of the radar system.

Abstract: A radar system, apparatus, architecture, and method are provided with a transmitter that produces a plurality of distinct FanTOM signals that are transmitted as N RF-encoded transmit signals in an overlapped fashion such that the pulse repetition interval and frame length are kept short; a receiver that processes target return signals reflected from the N RF-encoded transmit signals with a mixer to produce an IF signal which is filtered with one or more notch filters clocked with a sampling clock frequency to control harmonic notch frequencies to suppress transmitter spill-over and close-in self-clutter interference, thereby producing a filtered IF signal that is converted to a digital signal with an analog-to-digital converter that is clocked with the sampling clock frequency; and a radar processor that processes the digital signal to generate a range spectrum comprising N segments that correspond, respectively, to the N RF-encoded transmit signals.

Abstract: A distributed aperture radar system, apparatus, architecture, and method are provided for generating a completed virtual array aperture by using a radar control processing unit to construct a sparse forward-backward matrix from one or more sparse measurement array vectors and to generate beamforming outputs of a completed virtual array aperture by performing forward-backward matrix completion processing of the sparse forward-backward matrix to construct a completed forward-backward matrix by filling holes in the sparse MIMO virtual array aperture to suppress spurious sidelobes caused by holes in the sparse forward-backward matrix, thereby enabling computation of one or more super resolution direction of arrival (DoA) estimation values based on the completed forward-backward matrix.

Abstract: A signal processing system and method includes a first input configured to receive an input signal range profile. The input signal range profile includes near-range interference signals. A second input is configured to receive a reference signal range profile; and a processor is configured to perform steps including: executing a recursive least squares operation to determine coefficient values of a finite impulse response (FIR) filter, wherein the coefficient values are selected to minimize a difference between the input signal range profile and the reference signal range profile when the reference signal range profile is filtered through the FIR filter to generate a filtered reference signal range profile, and subtracting the filtered reference signal range profile from the input signal range profile to remove the near-range interference signals from the input signal range profile.

Abstract: In one example, a radar circuit uses computer processing circuitry for processing data corresponding to reflection signals via a sparse array. Output data indicative of signal magnitude associated with the reflection signals is generated, and then angle-of-arrival information is discerned therefrom by (e.g., iteratively): correlating the output data with at least one spatial frequency support vector indicative of a correlation peak for the output data; generating upper-side and lower-side support vectors which are neighbors along the spatial frequency spectrum for said at least one spatial frequency support vector, and providing, via a correlation of the upper-side and lower-side support vectors and said at least one spatial frequency support vector, at least one new vector that is more refined along the spatial frequency spectrum for said at least one spatial frequency support vector.

Abstract: Exemplary aspects are directed to a radar-based detection circuit or system with signal reception circuitry to receive reflection signals in response to radar signals transmitted towards objects. The system may include logic/computer circuitry and a multi-input multi-output (MIMO) virtual array to enhance resolution or remove ambiguities otherwise present in processed reflection signals. The MIMO array may include sparse linear arrays, each being associated with a unique antenna-element spacing from among a set of unique co-prime antenna-element spacings.

Abstract: A radar system, apparatus, architecture, and method are provided for generating a difference co-array virtual aperture by using a radar control processing unit to coherently combine virtual array apertures from multiple small aperture radar devices to construct a sparse MIMO virtual array aperture and to construct an extended difference co-array virtual array aperture that is larger than the MIMO virtual array aperture by using an FFT hardware accelerator to perform spectral-domain auto-correlation based processing of the sparse MIMO virtual array aperture to fill in holes in the sparse MIMO virtual array aperture and to suppress spurious sidelobes caused by holes in the sparse MIMO virtual array aperture.

Abstract: Aspects of the present disclosure relate to a radar system and method of operation thereof, whereby a spectrogram is generated by processing circuitry of a radar system by converting samples, of reflections of transmitted radar signals reflected by an object in an environment of the radar system, into a time-frequency domain, determining at least one threshold based on at least one MIN-of-MAX value for magnitudes of a frequency signal of the spectrogram, generating an interference-suppressed spectrogram by removing or attenuating interference components from the spectrogram based on the at least one threshold, and generating interference-suppressed samples based on the interference-suppressed spectrogram.

Abstract: Aspects of the present disclosure are directed toward apparatuses and/or methods involving the communication of radar signals. Certain aspects involve communicating time division multiplexing (TDM) multi-input multi-output (MIMO) radar signals, having pulses with a chirp interval time (CIT) that is different for respective chirps. Positional characteristics of a target may be ascertained based upon both the CIT between each chirp in the communicated radar signals and the time between each corresponding chirp in received ones of the signals reflected by the target. Communication of the radar signals may involve utilizing a combination of antennas to provide a virtual aperture.

Abstract: A method and system are provided to resolve Doppler ambiguity and multiple-input, multiple-output array phase compensation issues present in Time Division Multiplexing MIMO radars by estimating an unambiguous radial velocity measurement. Embodiments apply a disambiguation algorithm that dealiases the Doppler spectrum to resolve the Doppler ambiguity of a range-Doppler detection. Phase compensation is then applied for corrected reconstruction of the MIMO array measurements. The dealiasing processing first forms multiple hypotheses associated with the phase corrections for the radar transmitters based on a measured radial velocity of a range-Doppler cell being processed. A correct hypothesis, from the multiple hypotheses, is selected based on a least-spurious spectrum criterion. Using this approach, embodiments require only single-frame processing and can be applied to two or more transmitters in a TDM MIMO radar system.

Abstract: A radar system utilizing a linear chirp that can achieve a larger MIMO virtual array than traditional systems is provided. Transmit channels transmit distinct chirp signals in an overlapped fashion such that the pulse repetition interval is kept short and the frame is kept short. This alleviates range migration and aids in achieving a high frame update rate. The chirp signals from differing transmitters can be separated on receive in the range spectrum domain, such that a MIMO virtual array construction is possible. Distinct chirps are delayed versions of the first chirp signal. Chirps overlap in the fast-time domain, but due to delay, there is separation in the range spectrum domain. When the delay is at least the instrument round-trip delay, transmitters are separable. Further, the wavelengths are identical across transmitters such that there is no residual-range versus angle ambiguity issue present in the claimed frequency-offset modulation range division MIMO system.

Abstract: Aspects of the present disclosure are directed to radar signal processing apparatuses and methods. As may be implemented in accordance with one or more embodiments, digital signals representative of received reflections of radar signals transmitted towards a target are mathematically processed to provide or construct a matrix pencil based on or as a function of a forward-backward matrix. Eigenvalues of the matrix pencil are computed and an estimation of the direction of arrival (DoA) of the target is output based on the computed eigenvalues.

Abstract: A linear chirp radar system, apparatus and method use a radar control processing unit to control an LFM radar front end which generates analog-to-digital (ADC) sample signals from one or more target return signals received in response to transmitted linear chirp radar signals, where the radar control processing unit is connected and configured to mitigate range migration by directly filtering the ADC samples using a modified Doppler filter that is tuned to fast-time scaled, slow-time frequencies to generate a focused ADC Doppler cube, and by applying a Fourier Transform on each Doppler cell in the focused ADC Doppler cube to generate a focused range-Doppler cube.

Abstract: A vehicle radar system, apparatus and method use a radar control processing unit to generate a target response signal in at least a first dimension from compressed radar data signals and to perform cell-averaging constant false alarm rate (CA-CFAR) target detection by convolving the target response signal with a weighted kernel window signal in a frequency domain using a Fast Fourier Transform hardware accelerator, an element-wise multiplier, and an Inverse Fast Fourier Transform hardware accelerator to generate an output signal having a sign that indicates a target detection decision.

Abstract: A radar receiver comprising: an ADC (510) that samples analogue intermediate frequency, IF, signalling in order to generate digital signalling, wherein the digital signalling comprises a plurality of digital-values; a digital processor that populates a 2-dimensional array of bin-values based on the digital-values, such that: a first axis of the 2-dimensional array is a fast time axis and a second axis of the 2-dimensional array is a slow time axis; and a sampling-rate-adjuster that is configured to set a sampling rate associated with the bin-values in the 2-dimensional array based on an index of the slow time axis. The digital processor also performs DFT calculations on the bin-values in the 2-dimensional array along the fast time axis and the slow time axis in order to determine the range and velocity of any detected objects.