Radar data compression using hybrid spectral transformation

Abstract

Methods and systems for efficient radar data compression and reconstruction are provided. A radar processing unit receives and digitizes radar signals reflected from one or more objects, stores a first set of band indexed samples, while omitting a second set of band indexed samples from the storage. A plurality of sets of cross-correlation values are calculated between the second set of samples and the first set of samples within a local sliding window context. These values are used to reconstruct the second set of samples. The reconstructed and stored samples are then utilized to generate a radar map (e.g., a distance-velocity map, a distance-Doppler map, or a distance-angle map, which depends on the radar cube compression axis).

Description

TECHNICAL FIELD

[0001] The present disclosure relates to radar data compression methods, systems, and non-transitory computer-readable media using hybrid spectral transforms. BACKGROUND

[0002] In modern radar systems, including automotive radar systems, accurate detection and tracking of objects is critical for various applications including driver assistance systems and autonomous driving. These systems rely on radar sensors to capture and process large amounts of data about the surrounding environment, such as range, velocity, and angle information. The collected data is often referred to as a radar data cube.

[0003] Efficient storage and processing of radar data cubes is a significant challenge associated with these radar systems. The large amount of data produced requires a large amount of memory resources. As automotive radar system-on-chip (SoC) integrates more antennas and increases its data resolution, the demand for memory increases proportionally, leading to higher cost and power consumption. Furthermore, key information for accurate direction of arrival (DoA) estimation and velocity measurement, including phase information for one or more signals, needs to be preserved, further complicating the data handling process.

[0004] Current practices for radar data storage typically involve storing the entire radar data cube, including all fast-time samples and slow-time samples. While this approach ensures that all relevant information is preserved, it also results in a large memory consumption. Various data compression techniques have been explored to address this issue, including lossy and lossless methods. However, these methods often face limitations, such as inability to preserve phase information or requiring complex reconstruction processes. SUMMARY

[0005] A method includes:

[0006] receiving radar signals reflected from one or more objects;

[0007] sampling the received radar signals to produce a first set of band indexed samples and a second set of band indexed samples;

[0008] storing the first set of band indexed samples at a computer-readable memory device;

[0009] omitting the second set of band indexed samples from the computer-readable memory device;

[0010] computing and storing a plurality of cross-correlation values for the second set of band indexed samples;

[0011] reconstructing the second set of band indexed samples using the stored plurality of cross-correlation values and the stored first set of band indexed samples; and

[0012] generating a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

[0013] According to one or more embodiments, computing a set of cross-correlation values for a respective band indexed sample of the second set of band indexed samples comprises computing cross-correlation values between the band indexed sample and a subset of band indexed samples of the first set of band indexed samples that are adjacent to the respective band indexed sample.

[0014] According to one or more embodiments, computing and storing the plurality of sets of cross-correlation values comprises selecting a local sliding window context for the respective band indexed sample of the second set of band indexed samples, the local sliding window context comprising the subset of band indexed samples of the first set of band indexed samples that are adjacent to the respective band indexed sample.

[0015] According to one or more embodiments, each band indexed sample corresponds to a particular time instance along a slow time axis, and wherein generating the radar plot comprises generating a range-velocity plot.

[0016] According to one or more embodiments, the first set of band indexed samples comprises band odd indexed samples from the received radar signal, and wherein the second set of band indexed samples comprises band even indexed samples from the received radar signal.

[0017] According to one or more embodiments, the method is performed by a system-on-a-chip (SoC) integrated circuit system device comprising the computer readable memory device, wherein storing the first set of band indexed samples comprises storing the first set of band indexed samples in a radar data cube storage of the computer readable memory device, and wherein storing the plurality of sets of cross-correlation values comprises storing the plurality of sets of cross-correlation values in the radar data cube storage of the computer readable memory device.

[0018] According to one or more embodiments, reconstructing the second set of band indexed samples comprises solving a system of equations characterized by a Toeplitz matrix based on the plurality of sets of cross-correlation values and the first set of band indexed samples.

[0019] A radar system comprising:

[0020] a computer readable memory device;

[0021] a transceiver configured to:

[0022] receive radar signals reflected from one or more objects; and

[0023] sample the received radar signals to generate a first set of band indexed samples and a second set of band indexed samples; and

[0024] a radar processing unit configured to:

[0025] store, at the computer-readable memory device, the first set of band indexed samples;

[0026] omit, at the computer-readable memory device, storing the second set of band indexed samples;

[0027] compute and store a plurality of sets of cross-correlation values for the second set of band indexed samples;

[0028] reconstruct the second set of band indexed samples using the stored plurality of sets of cross-correlation values and the stored first set of band indexed samples; and

[0029] generate a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

[0030] According to one or more embodiments, to compute one of the plurality of sets of cross-correlation values for a band indexed sample in the second set of band indexed samples, the radar processing unit is configured to compute cross-correlation values between the band indexed sample in the second set of band indexed samples and a subset of band indexed samples in the first set of band indexed samples adjacent to the band indexed sample.

[0031] According to one or more embodiments, each band indexed sample corresponds to a particular time instance along a slow time axis, and wherein the radar plot comprises a range-velocity plot.

[0032] According to one or more embodiments, to compute and store the one set of cross-correlation values, the processing unit is further configured to:

[0033] for the band indexed sample in the second set of band indexed samples, select a local sliding window context comprising a subset of band indexed samples in the first set of band indexed samples adjacent to the band indexed sample.

[0034] According to one or more embodiments, the first set of band indexed samples comprises band odd indexed samples from the received radar signal, and wherein the second set of band indexed samples comprises band even indexed samples from the received radar signal.

[0035] According to one or more embodiments, the radar system further comprises a system on a chip (SoC) comprising the transceiver, the radar processing unit, and the computer-readable memory device.

[0036] According to one or more embodiments, to reconstruct the second set of band indexed samples, the radar processing unit is configured to solve a system of equations characterized by a Toeplitz matrix based on the plurality of sets of cross-correlation values and the first set of band indexed samples.

[0037] A non-transitory computer-readable medium storing a set of executable instructions that, when executed by one or more processors, manipulate the one or more processors to:

[0038] receive radar signals reflected from one or more objects;

[0039] sample the received radar signals to produce a first set of band indexed samples and a second set of band indexed samples;

[0040] store the first set of band indexed samples at a computer-readable memory device;

[0041] omit the second set of band indexed samples from the computer-readable memory device;

[0042] compute and store a plurality of sets of cross-correlation values for the second set of band indexed samples;

[0043] reconstruct the second set of band indexed samples using the stored plurality of sets of cross-correlation values and the stored first set of band indexed samples; and

[0044] produce a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

[0045] According to one or more embodiments, computing a set of cross-correlation values for a respective band indexed sample in the second set of band indexed samples includes computing cross-correlation values between the band indexed sample and a subset of band indexed samples in the first set of band indexed samples that are adjacent to the respective band indexed sample.

[0046] According to one or more embodiments, the first set of band indexed samples includes band odd indexed samples from the received radar signals, and wherein the second set of band indexed samples includes band even indexed samples from the received radar signals.

[0047] According to one or more embodiments, each band indexed sample corresponds to a particular time instance along a slow time axis, and wherein producing the radar plot includes producing a range-velocity plot.

[0048] According to one or more embodiments, computing and storing the plurality of cross-correlation values comprises selecting, for the respective band-indexed sample, a local sliding window context for the respective band-indexed sample, the local sliding window context comprising a sub-group of band-indexed samples in the first group of band-indexed samples that are adjacent to the respective band-indexed sample.

[0049] According to one or more embodiments, reconstructing the second group of band-indexed samples comprises solving a system of equations characterized by a Toeplitz matrix based on the plurality of cross-correlation values and the first group of band-indexed samples. BRIEF DESCRIPTION OF DRAWINGS

[0050] The disclosure can be better understood, and its numerous features and advantages can become apparent to those skilled in the art by reference to the following drawings, in which:

[0051] Figure 1 A frequency-modulated continuous wave (FMCW) radar system is shown.

[0052] Figure 2 Aspects of data compression techniques are shown in accordance with one or more embodiments.

[0053] Figure 3 A high-level block diagram of a radar system-on-a-chip (SoC) is shown in accordance with some embodiments.

[0054] Figure 4-1 And 4-2 A single-antenna range-velocity plot is depicted in accordance with some embodiments, showing range and velocity information derived from radar signal processing.

[0055] Figure 5 An operational routine for compressing radar data and generating a range-velocity plot based on the compressed data is depicted in accordance with some embodiments. DETAILED DESCRIPTION

[0056] A radar datacube is typically formed by sampling a radar signal in multiple dimensions: range (fast-time), velocity (slow-time), and space (antenna array). Fast-time sampling captures the time delay of the reflected signal, providing range information. Slow-time sampling involves capturing data over multiple radar pulses or chirps, enabling velocity estimation through Doppler processing. The spatial dimension is determined by signals received from multiple antenna elements, allowing for direction-of-arrival (DoA) estimation.

[0057] Embodiments of the technology described herein reduce the memory footprint of a radar datacube while maintaining the integrity of essential information, balancing memory usage, computational complexity, and accuracy of reconstructed data. Such embodiments enable deployment of high-resolution radar systems in cost-sensitive and power-constrained applications, including automotive applications, and can be used in a variety of radar signal scenarios. As non-limiting examples, the technology described herein can be used to reduce the memory footprint of a radar datacube in FMCW systems, phase-modulated continuous wave (PMC W) systems, pulse-based systems, spread-spectrum systems, orthogonal frequency-division multiplexing (OFDM) systems, ultra-wideband systems, and the like.

[0058] In certain embodiments, data compression is utilized to reduce the memory footprint of a radar datacube while preserving essential signal characteristics, such as phase information. For example, the technology described herein involves storing subsets of slow-time samples (such as band even indexed samples, band odd indexed samples, or other subsets) and using cross-correlation techniques to reconstruct omitted (e.g., band odd indexed) samples. By utilizing these techniques, the described embodiments balance memory efficiency and data integrity, facilitating more efficient processing and storage of radar data in automotive and other radar systems. The disclosed embodiments can also be applicable to different radar system configurations, and can be applied in a variety of scenarios where reduced data storage and computational efficiency is desired. Moreover, it should be appreciated that while various techniques are described herein with respect to data samples indexed along the slow-time axis of a radar datacube, such techniques can also be performed with respect to the fast-time axis and / or spatial axis in various embodiments and scenarios.

[0059] Figure 1 A frequency-modulated continuous wave (FMCW) radar system 100 is shown, which includes component circuitry for generating, transmitting, receiving, and processing radar signals. The FMCW radar system 100 is designed to detect and measure various properties of a target object 101. In one or more embodiments, the FMCW radar system 100 is an automotive radar system, which can be implemented as part of a vehicle and can be configured for use in one or more civilian automotive applications.

[0060] The radar system 100 includes a transmit section 110 and a receive section 150. The transmit section 110 includes a voltage-controlled oscillator (VCO) 112, which generates a highly linear frequency ramp signal 120. This ramp signal 120 is amplified by a power amplifier 114 and transmitted toward the target object 101 through a transmit antenna 116.

[0061] When the transmitted signal propagates and reflects off the target object 101, it returns to the radar system 100, where it is received by the receive antenna 152. The received signal then passes through a low noise amplifier (LNA) 154, which amplifies the weak signal. The amplified signal then mixes with the transmitted signal in a mixer 156 to produce an intermediate frequency (IF) signal, also known as a beat signal.

[0062] A feedback path 158 is included in the system to provide a portion of the transmitted signal from the transmission section 110 to the mixer 156. This feedback mechanism allows the transmitted signal to be compared with the received signal, facilitating the generation of the IF signal. The beat frequency, which is proportional to the distance to the target object 101, is extracted by this process.

[0063] The IF signal is further processed by an intermediate frequency amplifier (IFA) 160, which removes high frequency noise from the signal, and a low pass filter (LPF) 162, which removes unwanted high frequency components from the signal. The filtered signal is then digitized by an analog-to-digital converter (ADC) 164, converting the analog signal to digital form for subsequent processing.

[0064] The digitized signal 190 undergoes a fast Fourier transform (FFT) 195 to convert it from the time domain to the frequency domain, enabling the detection of objects at different distances and velocities represented as peaks in the resulting frequency spectrum. Distance information is then derived from the FFT output (199), and the strength of the reflected signal is used to determine the characteristics of the target object 101.

[0065] Figure 2 Aspects of a data compression technique used in one or more embodiments are shown, which advantageously exploit the relationship between the even indexed samples (e.g., 208, 210, 212, 214, 216, 218, 220) and the odd indexed samples (e.g., 209, 211, 213, 215, 217, 219, 221) along the slow time axis. The data samples 205 represent a sequence of data points obtained from radar signal processing, such that each data point corresponds to a particular time instance.

[0066] In the illustrated embodiment, the even indexed samples, represented by squares, are retained and stored, while the odd indexed samples, represented by circles, are omitted from storage. The selection of these samples follows a periodic pattern, where each stored even indexed sample is followed by an omitted odd indexed sample.

[0067] The cross-correlation 201 is performed between each skipped odd-indexed sample and its stored even-indexed sample's local context. For example, the skipped sample 215 is correlated with six adjacent even-indexed samples within a local sliding window, including samples 210, 212, 214, 216, 218, and 220. This use of a sliding window context helps capture the relationship between the skipped and stored samples over a small range of the signal, while each skipped sample only stores six cross-correlation values, significantly reducing the storage space required for those even-indexed samples. In certain embodiments, the cross-correlation values are recorded and utilized in a subsequent reconstruction step, allowing the skipped samples to be estimated based on the stored samples.

[0068] In this context, the use of cross-correlation provides a significant advantage over the use of autocorrelation. Cross-correlation measures the similarity between two different signals—in this case, the skipped odd-indexed sample and the surrounding stored even-indexed sample. This technique preserves the phase information of the signal, which enables the radar system incorporated to accurately determine parameters such as direction of arrival (DoA) and velocity. In contrast, autocorrelation involves comparing a signal to a delayed version of itself. While autocorrelation can reduce the number of data points required for storage, it tends to lose phase information because it involves the multiplication of a signal with its complex conjugate. This loss of phase information can lead to inaccuracies in the reconstructed data, particularly in applications where phase plays a role in the interpretation of the signal, such as in DoA estimation and Doppler processing.

[0069] As a non-limiting example, a set of neighboring even-indexed samples 210, 212, 214, 216, 218, and 220 are used to predict the value of a skipped odd-indexed sample, e.g., sample 215. The prediction is achieved by applying a normal equation method within the sliding window context of those samples. In the normal equation method, a matrix-vector equation is constructed, where the left-hand side matrix consists of auto-correlation values of the stored even-indexed slow-time samples, e.g., samples 210, 212, 214, 216, 218, and 220. This matrix typically takes the form of a Toeplitz matrix, which is characterized by constant diagonals. A Toeplitz matrix is a type of matrix where each diagonal from the top left to the bottom right is constant. This means that all elements along a diagonal have the same value. Formally, a matrix T is called a Toeplitz matrix if for all valid indices i and j, T[i,j] = T[i+1,j+1], so that the value of a matrix element depends only on the difference between column index and row index. The Toeplitz structure is advantageous because it allows the application of well-known complexity reduction techniques, such as the Levinson-Durbin algorithm, to efficiently solve the matrix-vector equation.

[0070] In certain embodiments, the right-hand side of this matrix-vector equation includes cross-correlation values between the skipped odd-indexed sample (e.g., 215) and the stored even-indexed samples (samples 210, 212, 214, 216, 218, and 220) from the same local sliding window context. Solving this system of equations yields linear prediction coefficients, which are then used to weight the even-indexed samples for predicting the value of the skipped sample. This solution ensures that the reconstructed signal maintains high fidelity to the original, including phase information.

[0071] In certain embodiments, a separate mean square error (MSE) prediction vector is generated for each column of odd-indexed sub-sequences within the data cube, which corresponds to a different distance index. The area required for the cross-correlation computation and the MSE prediction decompression computation is relatively small compared to the area required for the original memory required to store all sample data (even-indexed samples and odd-indexed samples). Thus, the overall integrated circuit (IC) area savings can be quite large. Alternatively, if the original memory area is preserved, the described techniques enable storing approximately twice the number of slow-time samples, effectively doubling the capacity of the data cube without increasing the physical memory size.

[0072] Figure 3A high-level block diagram of a radar system-on-a-chip (SoC) 301 is shown, which integrates various circuit system components for radar signal processing and system management. In this and other embodiments, the techniques described herein leverage the integrated design of the radar SoC 301 to efficiently manage radar data.

[0073] In the depicted embodiment, the radar SoC 301 includes a radar processing unit 305 and a radar transceiver 310, along with communicatively coupled peripheral interfaces and power management circuitry 350 to the side. The radar processing unit 305 is configured to handle data processing and control functions. In various embodiments, the radar processing unit 305 includes one or more microcontroller units (MCUs), central processing units (CPUs), digital signal processors (DSPs), application specific integrated circuits (ASICs), and / or other types of hardware processors. Within the radar processing unit 305, a radar processing platform 320 encompasses the radar signal processing tools and hardware necessary to manage radar data. Notably, a substantial portion of the radar processing platform 320 is allocated to a radar data cube storage 322 (sometimes referred to herein as a “computer readable memory device 322”). The radar data cube storage 322 corresponds to a memory storage area designated for storing a radar data cube, which includes processed radar signals over multiple dimensions such as range, velocity, and angle. For example, the radar data cube storage 322 can store a range-Doppler antenna cube generated by the radar processing unit 305 by performing range compression (e.g., fast-time FFT) and Doppler compression (e.g., slow-time FFT) on samples (e.g., ADC samples) of reflected radar signals received via the antenna 315. In one or more embodiments, a large-scale IC area can be dedicated to the radar data cube storage, as this data typically needs to be efficiently managed and stored given its significant memory requirements. In one or more embodiments, as a non-limiting example, the radar data cube storage 322 can include random access memory (RAM), such as SRAM.

[0074] While in this example, the radar processing unit 305 and the radar transceiver 310 are shown as included in the same radar SoC 301, it should be understood that this arrangement is intended to be illustrative and not limiting. For example, in one or more other embodiments, the radar processing unit 305 and the radar transceiver 310 are instead implemented as separate components that are communicatively and electrically coupled together and that are not part of the same SoC. While in this example, the radar datacube storage 322 is shown as implemented as part of the radar processing unit 305, it should be understood that this is illustrative and not limiting. For example, in one or more other embodiments, the radar datacube storage 322 can be implemented via a memory or other computer-readable data storage coupled to the radar processing unit.

[0075] In the depicted embodiment, the radar processing unit 305 hosts a CPU platform 324 that manages execution of software and control algorithms for radar operations and overall system functionality. Connection circuitry 326 within the radar processing unit facilitates communication between different parts of the system and external interfaces (e.g., vehicle communication interface 340, high-speed data interface 344, and onboard Ethernet interface 348), supporting various communication protocols and interfaces.

[0076] The radar transceiver 310 section is responsible for signal generation, transmission, reception, and initial signal conditioning. Within the radar transceiver 310, signal generation and transmission circuitry 312 generates radar signals for transmission through an antenna array. Signal conditioning and digitization circuitry 314 captures reflected radar signals, performs signal conditioning, and digitizes received signals for further processing. In one or more embodiments, the signal conditioning and digitization circuitry 314 includes at least one analog-to-digital converter (ADC) that samples and digitizes reflected radar signals to produce digital ADC samples representing the reflected radar signals.

[0077] The functional safety circuitry block 316 ensures that the system operates safely even in the presence of faults or failures by monitoring system health and implementing necessary safety measures to prevent hazardous situations. In various embodiments, the functional safety circuitry block 316 provides various additional state management, such as compliance with one or more relevant safety standards (e.g., ISO 26262 for automotive systems), and / or prevention of false detections or missed detections from affecting vehicle control decisions.

[0078] The radar SoC 301 interfaces with external components through several interfaces. A vehicle communication interface 340 allows communication with the vehicle’s internal network, such as a controller area network (CAN) and other automotive communication protocols. A high-speed data interface 344 provides a high-speed communication link for data transfer between the radar system and other vehicle systems. An onboard Ethernet 348 facilitates data communication over Ethernet, supporting high-bandwidth data transfer requirements.

[0079] Power management within the radar SoC 301 is handled by power management circuitry 350, which regulates and manages power distribution to ensure efficient and reliable operation.

[0080] The radar transceiver 310 interfaces with multiple antenna arrays. Transmit antennas 313 are used to transmit radar signals generated by signal generation and transmission circuitry 312. Receive antennas 315 are used to receive reflected radar signals, which are then processed by signal conditioning and digitization circuitry 314.

[0081] Figure 4-1 And Figure 4-2 Single antenna range-velocity plots are depicted, showing range and velocity information derived from radar signal processing. These plots provide a visual representation of radar data in a two-dimensional plot, with the horizontal axis representing range (r) and the vertical axis representing velocity (v). The graduated scale on the right side of each plot indicates the strength or magnitude of the radar signal, with different shades of color representing different levels of signal strength.

[0082] Figure 4-1 A range-velocity plot 410 is shown, obtained without applying data compression techniques. The range-velocity plot 410 serves as a baseline, presenting the full radar data as captured by a single antenna system. The continuous distribution of intensity across the range-velocity plot 410 indicates the presence of multiple detected objects or targets, with variations in range and velocity. The variations correspond to different signal intensities, which are related to the range and reflectivity of the detected objects.

[0083] Figure 4-2 A range-velocity plot 420 is presented, after applying the data compression and subsequent decompression techniques described in this disclosure to the range-velocity plot 410. The range-velocity plot 420 demonstrates the ability to reconstruct the radar data cube after such compression, allowing key information about the targets’ range and velocity to be preserved. Figure 4-1 A comparison between Figure 4-2 highlights the effectiveness of the compression techniques in maintaining the fidelity of the original radar data while reducing storage requirements. Compared to Figure 4-1 , Figure 4-2 the features preserved in show the preservation of characteristics such as target position and velocity after decompression, without significant loss of detail.

[0084] Figure 5 An operational routine 500 is depicted for compressing radar data and generating a range-velocity map based on the compressed data. The operational routine 500 can be performed by a radar system, e.g., a radar SoC system 301 similar to Figure 3 .

[0085] The operational routine 500 begins in block 505, where the radar system receives radar signals reflected from one or more target objects. The received radar signals include information about the location and velocity of the target objects. In certain embodiments, the reception of such radar signals includes one or more pre-processing stages, e.g., removal of noise, removal of unwanted frequency components, digitization, and / or transformation (e.g., converting the received signals from time domain to frequency domain). The routine proceeds to block 510.

[0086] In block 510, a first set of band-indexed samples from the received (and possibly pre-processed) radar signals is stored in, e.g., a radar data-cube storage associated with the radar system (e.g., the radar data-cube storage 322 of the radar SoC system 301). In the depicted embodiment, such samples correspond to a particular time instance along the slow-time axis, and are selected according to a predetermined pattern. The band-indexed samples stored in this step form the basis for subsequent computations and data reconstruction. The routine proceeds to block 515. Figure 3

[0087] In block 515, the radar system stores a second set of band-indexed samples from the received radar signals. These stored samples also correspond to a particular time instance along the slow-time axis, but are not stored directly. Instead, their information is later reconstructed based on the stored samples and computed cross-correlation values. The routine proceeds to block 520.

[0088] In block 520, the radar system computes and stores a set of cross-correlation values for each omitted band-indexed sample in the second set of band-indexed samples. In the depicted embodiment, these cross-correlation values are computed between the omitted sample and a plurality of neighboring band-indexed samples from the first set of band-indexed samples, within a local sliding window context. The cross-correlation values capture the relationship between the omitted sample and the stored samples, enabling the later reconstruction of the omitted sample. The routine proceeds to block 525.

[0089] In block 525, the omitted band-indexed samples in the second set of band-indexed samples are reconstructed using the stored cross-correlation values and the stored band-indexed samples in the first set of band-indexed samples. The reconstruction process involves solving a system of equations characterized by a Toeplitz matrix, using the cross-correlation values and the stored samples. This enables the accurate estimation of the omitted samples. The routine proceeds to block 530.​

[0090] At block 530, a range-velocity map is generated based at least in part on the stored samples in the first set of samples and the reconstructed samples in the second set of samples. The range-velocity map visually represents range and velocity information for the detected target object, providing a comprehensive overview of the target object's position and motion. It will be appreciated that in embodiments in which samples are collected and omitted along the fast time axis rather than the slow time axis of the radar data cube, the radar system generates a range-Doppler map (rather than a range-velocity map) as a result of the reconstruction of such samples, providing a two-dimensional representation of the target object's range and relative velocity. Similarly, in embodiments in which samples are collected and omitted along the spatial axis, the radar system generates a range-angle map, providing a two-dimensional representation of the object's range and its angular position.

[0091] In some embodiments, certain aspects of the techniques described above can be implemented by one or more processors executing software. The software includes one or more sets of instructions executable by the one or more processors. The software can include instructions and certain data that, when executed by the one or more processors, manipulate the one or more processors to perform one or more aspects of the techniques described above. The non-transitory computer-readable storage medium can include, for example, a magnetic or optical disk storage such as a compact disk (CD) or digital versatile disk (DVD), a solid state memory like flash memory, a cache, random access memory (RAM) or any other storage medium(s) that can be used to store desired program code means in the form of computer-executable instructions or data structures and that can be accessed by a general purpose or special purpose computer system. The software can be distributed on such a computer-readable storage medium, or can be distributed to a user of a computer system on a basis such as through the Internet, an intranet, an extranet, or a combination of tangible and / or non-transitory computer-readable storage media and computer networks. Here, computer-readable storage medium generally can include tangible computer- readable media that is non-transitory storage media that stores data for block-based access by a machine. Examples of a computer-readable storage medium that stores for block-based access by a machine can include magnetic, optical, or semiconductor storages such as magnetic disks, optical discs, and flash memory, as well as forms of cache or RAM. Computer-readable storage media can also include non-tangible media such as electrical, optical, acoustical, or other form of propagated signals - such as carrier waves, infrared signals, digital signals, etc. Examples of a computer-readable medium that stores for block-based access by a machine can include magnetic, optical, or semiconductor storages such as magnetic disks, optical discs, and flash memory, as well as forms of cache or RAM. Computer-readable storage media can also include non-tangible media such as electrical, optical, acoustical, or other form of propagated signals - such as carrier waves, infrared signals, digital signals, etc.

[0092] A computer-readable storage medium can include any storage media, or combination of storage media, accessible by a computer system during use to provide instructions and / or data to the computer system. Such storage media can include, but is not limited to, optical media (e.g., compact disc (CD), or Blu-Ray disc), magnetic media (e.g., floppy disc, or magnetic hard drive), volatile memory (e.g., random access memory (RAM) or cache), non-volatile memory (e.g., read-only memory (ROM) or Flash memory), or microelectromechanical systems (MEMS)-based storage media. The computer-readable storage medium can be embedded in the computing system (e.g., system RAM or ROM), fixed in the computing system (e.g., a magnetic hard drive), or provided external to the computing system (e.g., an optical disc or a USB flash drive), or accessible via a wired or wireless network (e.g., a network accessible storage device).

[0093] It should be noted that the activities or elements described above in the general description are not all-inclusive and that specific activities or elements described can not be required, that additional activities or elements can be utilized, and that not all of the activities or elements described above need be utilized. Further, the order in which activities are listed is not necessarily the order in which the activities are performed. Also, the concepts have been described with reference to particular embodiments. Various modifications and changes can be made thereto without departing from the scope of the disclosure as set forth in the following claims. The disclosure and the figures are, therefore, to be regarded as illustrative rather than restrictive, and all such modifications are intended to be included within the scope of the present disclosure.

[0094] Benefits, other advantages, and solutions to problems have been described above with regard to specific embodiments. However, the benefits, advantages, solutions to problems, and any feature(s) that can cause any benefit, advantage, or solution to occur or become more pronounced are not to be construed as a critical, required, or essential feature of any or all the claims. Furthermore, the particular embodiments disclosed above are illustrative only as the disclosed subject matter can be modified and practiced in different but equivalent manners that are apparent to those skilled in the art having the benefit of the teachings herein. No limitations are intended to the details of construction or design herein shown, other than as described in the claims below. It is therefore evident that the particular embodiments disclosed above can be altered or modified and all such variations are considered within the scope of the disclosed subject matter. Accordingly, the protection sought herein is as set forth in the following claims.

Claims

1. A method characterized by, comprising: receiving radar signals reflected from one or more objects; sampling the received radar signals to produce a first set of band indexed samples and a second set of band indexed samples; storing the first set of band indexed samples at a computer readable memory device; omitting the second set of band indexed samples from the computer readable memory device; computing and storing a plurality of sets of cross-correlation values for the second set of band indexed samples; reconstructing the second set of band indexed samples using the stored plurality of sets of cross-correlation values and the stored first set of band indexed samples; and producing a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

2. The method of claim 1, wherein, Computing a set of cross-correlation values for a respective band indexed sample in the second set of band indexed samples includes computing cross-correlation values between the band indexed sample and a subset of band indexed samples in the first set of band indexed samples that are adjacent to the respective band indexed sample.

3. The method of claim 2, wherein, Computing and storing the plurality of sets of cross-correlation values includes selecting, for the respective band indexed sample in the second set of band indexed samples, a local sliding window context that includes the subset of band indexed samples in the first set of band indexed samples that are adjacent to the respective band indexed sample.

4. A radar system, characterized by comprising: a computer readable memory device; a transceiver configured to: receive radar signals reflected from one or more objects; and sample the received radar signals to produce a first set of band indexed samples and a second set of band indexed samples; and a radar processing unit configured to: store the first set of band indexed samples at the computer readable memory device; omit storing the second set of band indexed samples at the computer readable memory device; compute and store a plurality of sets of cross-correlation values for the second set of band indexed samples; reconstruct the second set of band indexed samples using the stored plurality of sets of cross-correlation values and the stored first set of band indexed samples; and produce a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

5. The radar system of claim 4, wherein, To compute a set of cross-correlation values for a band indexed sample in the second set of band indexed samples, the radar processing unit is configured to compute cross-correlation values between the band indexed sample in the second set of band indexed samples and a subset of band indexed samples in the first set of band indexed samples that are adjacent to the band indexed sample.

6. The radar system of claim 4, wherein, To compute and store the set of cross-correlation values, the processing unit is further configured to: select, for the band indexed sample in the second set of band indexed samples, a local sliding window context that includes the subset of band indexed samples in the first set of band indexed samples that are adjacent to the band indexed sample.

7. A non-transitory computer-readable medium, comprising: storing a set of executable instructions that, when executed by one or more processors, manipulate the one or more processors to: receive radar signals reflected from one or more objects; sample the received radar signals to produce a first set of band indexed samples and a second set of band indexed samples; store the first set of band indexed samples at a computer readable memory device; deriving a second set of band indexed samples from the first set of band indexed samples; computing and storing a plurality of sets of cross-correlation values for the second set of band indexed samples; reconstructing the second set of band indexed samples using the stored plurality of sets of cross-correlation values and the stored first set of band indexed samples; and generating a radar plot based at least in part on the reconstructed second set of band indexed samples and the stored first set of band indexed samples.

8. The non-transitory computer-readable medium of claim 7, wherein, Computing a set of cross-correlation values for a respective band indexed sample in the second set of band indexed samples includes computing cross-correlation values between the band indexed sample and a subset of band indexed samples in the first set of band indexed samples that are adjacent to the respective band indexed sample.

9. The non-transitory computer-readable medium of claim 7, wherein, Computing and storing the plurality of sets of cross-correlation values includes selecting, for the respective band indexed sample, a local sliding window context for the respective band indexed sample, the local sliding window context including the subset of band indexed samples in the first set of band indexed samples that are adjacent to the respective band indexed sample.

10. The non-transitory computer-readable medium of claim 7, wherein, Reconstructing the second set of band indexed samples includes solving a system of equations characterized by a Toeplitz matrix based on the plurality of sets of cross-correlation values and the first set of band indexed samples.

Images