Automotive Radar Systems Research Articles

Learning Approach to FMCW Radar Target Classification With Feature Extraction From Wave Physics

Target classification is of great value when the number of transceivers in a sensing system is relatively small. This article proposes a high-accuracy and efficient classification method with machine learning techniques on frequency-modulated-continuous-wave (FMCW) radar. We first establish the novel mapping relationship from physical space to range-Doppler (PS-RD) image based on wave propagation theory, by which four physical features that effectively capture the kinematic and geometrical characteristics of targets, including speed, total reflectivity (ToRe), area, and incidence angle, are extracted from range-Doppler (R-D) image. Then, a multilayer perceptron (MLP) with a single hidden layer is employed to realize the classification. Since the above-mentioned four physical features, derived from wave physics, are chosen as the input of the neural network, our classifier does not work in a black-box way. The computational complexity of the whole classifier is the same as that of a 2-D fast Fourier transform (FFT), which guarantees a real-time operation. As an example, the proposed classifier is applied to the automotive radar system, where road targets are to be classified into five categories, including pedestrian, bike, sedan, truck/bus, and other static objects. Real-world data obtained from 77 GHz FMCW radars are provided for validation, where the proposed physics-assisted classifier turns out to outperform the state of the art in automotive radar application. The overall accuracy of the real data is about 99% even with complex multiple-target cases.

Rational Design of WSe2 /WS2 /WSe2 Dual Junction Phototransistor Incorporating High Responsivity and Detectivity.

The excellent semiconducting properties and ultrathin morphological characteristics allow van der Waals (vdW) heterostructures based on 2D materials to be promising channel materials for the next-generation optoelectronic devices, especially in photodetectors. Although various 2D heterostructure-based photodetectors have been developed, the unavoidable trade-off between responsivity and detectivity remains a critical issue for these devices. Here, an ingenious phototransistor based on WSe2 /WS2 /WSe2 dual-vdW heterostructures is constructed, performing both high responsivity and detectivity. In the charge neutrality point (gate voltage of -15V and bias voltage of 1V), this device demonstrates a pronounced photosensitivity, accompanying with high detectivity of 1.9×1014 Jones, high responsivity of 35.4AW-1 , and fast rise/fall time of 3.2/2.5ms at 405nm with power density of 60µWcm-2 . Density functional theory calculations, energy band profiles, and optoelectronic characteristics jointly verify that the high performance is ascribed to the distinctive device design, which not only facilitates the separation of photogenerated carriers but also produces a strong photogating effect. As a feasible application, an automotive radar system is demonstrated, proving that the device has considerable potential for application in vehicle intelligent assisted driving.

A 77-GHz Low-Conversion-Loss Fourth-Harmonic Mixer on SISL Platform

This letter presents a 77-GHz low-conversion-loss (CL) fourth-harmonic mixer based on the substrate-integrated suspended line (SISL) platform. Based on the traditional stub-type harmonic mixing structure, we add several additional transmission lines, which will participate in the impedance matching of the diode together with stubs to reduce the CL. A 3-D electromagnetic (EM) diode model is established for accurate simulation at millimeter-wave frequency. The presented SISL mixer was fabricated by the low-cost printed circuit board (PCB) process and was self-packaged in FR4 substrates. When the local oscillator (LO) frequency is 19.45 GHz and the radio frequency (RF) is 78 GHz, the minimum CL is 12.2 dB under an LO input power of 12 dBm. This mixer is suitable for 77-GHz automotive anti-collision radar systems.

Computationally Efficient Long Binary Sequence Designs With Low Autocorrelation Sidelobes

The design of binary sequences with low autocorrelation sidelobes is critically important to diverse applications, including radar and sonar as well as communications. Many computational methods have been proposed recently to design binary sequences through optimizing the integrated sidelobe level (ISL), which is a commonly used optimization metric. However, due to the high computational complexities, few of them can be used for long binary sequence designs, which are needed in, for example, low-cost commercial automotive radar systems. To improve the performance of the recently introduced fast Fourier transform (FFT) based CANARY algorithm for long aperiodic binary sequence designs, we introduce a new penalty function into the ISL metric and also propose a new framework of double-loop iterations. These new ideas are also extended to design long periodic binary sequences with low autocorrelation sidelobes and long aperiodic and periodic binary sequences with low correlation zones. The associated cost functions are still optimized efficiently via FFT-based iterations, making the proposed algorithms computationally efficient and, hence, suitable for long binary sequence designs. Moreover, the convergence properties and the computational complexities of these algorithms are analyzed. Numerical examples are provided to demonstrate the effectiveness of the proposed algorithms.

Finite Impulse Response Filter-Based Track Formation for Preceding Vehicle Tracking Using Automotive Radars.

Automotive radars, which are used for preceding vehicle tracking, have attracted significant attention in recent years. However, the false measurements that occur in cluttered roadways hinders the tracking process in vehicles; thus, it is essential to develop automotive radar systems that are robust against false measurements. This study proposed a novel track formation algorithm to initialize the preceding vehicle tracking in automotive radar systems. The proposed algorithm is based on finite impulse response filtering, and exhibited significantly higher accuracy in highly cluttered environments than a conventional track formation algorithm. The excellent performance of the proposed algorithm was demonstrated using extensive simulations under real conditions.

Geometric Sequence Decomposition-Based Interference Cancellation in Automotive Radar Systems

Radar signals generated from other vehicles can act as interference to ego-vehicle, which degrades the inherent detection performance of radar. In this paper, we address an efficient solution to the interference problem in frequency-modulated continuous wave (FMCW)-based automotive radar systems, named geometric sequence decomposition based interference cancellation (GSD-IC). With the method of GSD-IC, we can decompose the received signal into different non-orthogonal superposed signals, which means the interference signal and the signal reflected from the desired target are separated. It is based on the facts that 1) each single sampled signal can be interpreted as a geometric sequence and 2) the useful physical features, such as the time delay and the Doppler frequency, are extracted after converting the superposition of these geometric sequences into several transformed matrices. Through this approach, we can achieve effective interference signal cancellation while minimizing the loss of meaningful target information. Moreover, the proposed method does not require the generation of specific radar waveforms, and can mitigate interference through signal processing even with existing waveforms. Numerical results show that our algorithm outperforms existing methods in various scenarios.

A Two-Stage Support Vector Machine and SqueezeNet System for Range-Angle and Range-Speed Estimation in a Cluttered Environment of Automotive MIMO Radar Systems

This paper proposes a two-stage deep-learning approach for frequency modulated continuous waveform multiple‐input multiple‐output (FMCW MIMO) radar embedded in cluttered and jammed environments. The first stage uses the support vector machine (SVM) as a feature extractor that discriminates targets from clutters and jammers. In the second stage, the angle, range, and Doppler estimations of the extracted targets are treated by the SqueezeNet deep convolutional neural network (DCNN) as a multilabel classification problem. The performance of the proposed hybrid SVM-SqueezeNet method is very close to the one achieved by the SqueezeNet only but with the advantage of identifying the type of targets and reducing the training time required by the SqueezeNet.

Automotive radar interference study for different radar waveform types

Mutual interference between different radar waveforms used in automotive radar applications is studied. The existing interference analysis is extended to a generalised radar-to-radar interference equation that covers most of the common interference scenarios for automotive radar systems. The outcome of the generalised equation is demonstrated for a number of typical scenarios where radars with different continuously transmitting waveforms are involved. The proposed equation can be used to characterise the received interference and its features by analysing the instantaneous beat frequency of the victim radar. Moreover, an interference analysis of phase-coded frequency-modulated continuous waveforms is performed and demonstrated experimentally by using real-time automotive radars for the first time in the literature. The experimental results corroborate the interference analysis of different waveforms and validate the proposed generalised interference equation under various conditions.

Object detection for automotive radar point clouds \u2013 a comparison

Automotive radar perception is an integral part of automated driving systems. Radar sensors benefit from their excellent robustness against adverse weather conditions such as snow, fog, or heavy rain. Despite the fact that machine-learning-based object detection is traditionally a camera-based domain, vast progress has been made for lidar sensors, and radar is also catching up. Recently, several new techniques for using machine learning algorithms towards the correct detection and classification of moving road users in automotive radar data have been introduced. However, most of them have not been compared to other methods or require next generation radar sensors which are far more advanced than current conventional automotive sensors. This article makes a thorough comparison of existing and novel radar object detection algorithms with some of the most successful candidates from the image and lidar domain. All experiments are conducted using a conventional automotive radar system. In addition to introducing all architectures, special attention is paid to the necessary point cloud preprocessing for all methods. By assessing all methods on a large and open real world data set, this evaluation provides the first representative algorithm comparison in this domain and outlines future research directions.

Radar Signal Decomposition in Steering Vector Space for Multi-Target Classification

In automotive frequency-modulated continuous wave radar systems, targets with similar ranges and velocities might be detected as a single target because their target information overlaps on the range-velocity (RV) plane. When the information of multiple targets overlaps on the RV plane, target classification cannot be performed accurately. Therefore, we propose a method to decompose the overlapped target information on the RV plane for effective target classification. To separate each target’s radar signal, we use an orthogonal projection, in which the steering vector formed from the angle information of the desired target is projected onto the kernel space composed of other steering vectors. To verify the performance of the proposed method, a convolutional neural network-based target classification is performed on the decomposed radar signals. Compared to the angle-fast Fourier transform (FFT)-based signal decomposition, the orthogonal projection-based signal decomposition shows higher classification accuracy, which implies that the target information overlapped on the RV plane is effectively separated. When information of two targets is overlapped, the classification accuracy of the proposed method is 5.9%p higher than that of the conventional method. Moreover, when the information of three targets is overlapped, the accuracy is improved by 25.5%p.

Compressive Sensing-Based SAR Image Reconstruction from Sparse Radar Sensor Data Acquisition in Automotive FMCW Radar System.

In this paper, we propose a method for reconstructing synthetic aperture radar (SAR) images by applying a compressive sensing (CS) technique to sparsely acquired radar sensor data. In general, SAR image reconstruction algorithms require radar sensor data acquired at regular spatial intervals. However, when the speed of the radar-equipped platform is not constant, it is difficult to consistently perform regular data acquisitions. Therefore, we used the CS-based signal recovery method to efficiently reconstruct SAR images even when regular data acquisition was not performed. In the proposed method, we used the l1-norm minimization to overcome the non-uniform data acquisition problem, which replaced the Fourier transform and inverse Fourier transform in the conventional SAR image reconstruction method. In addition, to reduce the phase distortion of the recovered signal, the proposed method was applied to each of the in-phase and quadrature components of the acquired radar sensor data. To evaluate the performance of the proposed method, we conducted experiments using an automotive frequency-modulated continuous wave radar sensor. Then, the quality of the SAR image reconstructed with data acquired at regular intervals was compared with the quality of images reconstructed with data acquired at non-uniform intervals. Using the proposed method, even if only 70% of the regularly acquired radar sensor data was used, a SAR image having a correlation of 0.83 could be reconstructed.

Mutual Interference Mitigation for Multiple Connected Automotive Radar Systems

The number of commercial and civilian vehicles equipped with automotive radars is expected to rise rapidly in the forthcoming years and with that comes the problem of increased mutual interference between the radar sensors, which can result in severely reduced radar sensitivity and increased false alarm rates. The difficulty and complexity of the problem increase with MIMO radar systems and multiply even further with a growing number of vehicles present on the scene. A system of connected vehicles can efficiently address this problem by sharing information amongst themselves. In this paper, we propose an efficient waveform design algorithm that seeks to minimize a collective cross-ambiguity function. Vehicles that can <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">talk</i> to each other, can perform the design online in a collaborative manner, or offline, in which case the radar codes can be designed and stored in a codebook for future use. The proposed coding scheme is computationally efficient for practical use and the incorporation of such schemes requires only a slight modification of the existing systems. Our numerical examples indicate that the proposed scheme can significantly reduce the interference power level in a desired area of the radar cross-ambiguity functions.

Weighted SPICE Algorithms for Range-Doppler Imaging Using One-Bit Automotive Radar

We consider the problem of range-Doppler imaging using one-bit automotive LFMCW1 or PMCW radar that utilizes one-bit ADC sampling with time-varying thresholds at the receiver. The one-bit sampling technique can significantly reduce the cost as well as the power consumption of automotive radar systems. We formulate the one-bit LFMCW/PMCW radar rangeDoppler imaging problem as one-bit sparse parameter estimation. The recently proposed hyperparameter-free (and hence user friendly) weighted SPICE algorithms, including SPICE, LIKES, SLIM and IAA, achieve excellent parameter estimation performance for data sampled with high precision. However, these algorithms cannot be used directly for one-bit data. In this paper we first present a regularized minimization algorithm, referred to as 1bSLIM, for accurate range-Doppler imaging using onebit radar systems. Then, we describe how to extend the SPICE, LIKES and IAA algorithms to the one-bit data case, and refer to these extensions as 1bSPICE, 1bLIKES and 1bIAA. These onebit hyperparameter-free algorithms are unified within the one-bit weighted SPICE framework. Moreover, efficient implementations of the aforementioned algorithms are investigated that rely heavily on the use of FFTs. Finally, both simulated and experimental examples are provided to demonstrate the effectiveness of the proposed algorithms for range-Doppler imaging using one-bit automotive radar systems.

Design of 77 GHz automotive angular radar array antenna with wide beamwidth and flat gain

A wide beam, flat gain planar array antenna based on beamforming technology is designed for automotive angular radar systems. The array antenna consists of three 1×10 series feed line arrays, which are designed by Dolph-Chebyshev antenna synthesis method with unequal spacing distribution to reduce the side lobes. In order to satisfy the needs of the automotive angle radar wide field of view (FOV), the beamforming techniques are used for the planar array antenna to broaden the beamwidth of the azimuth plane and improve the flatness of the gain. The simulation and the measured results agree well. The measured results show that the 1×10 line array sidelobe level is −19.7 dB; the 3×10 planar array antenna has a 3 dB beamwidth of 106.4°, and the azimuth gain flatness range can reach ± 45°, the maximum gain is 11.3 dBi.

Hybrid PDA/FIR Filtering for Preceding Vehicle Tracking Using Automotive Radars

This paper proposes a novel single vehicle tracking algorithm with enhanced reliability for automotive radar systems. The proposed algorithm overcomes the weaknesses of the probabilistic data association filter (PDAF) in single-target tracking in clutter. The PDAF is successful in normal situations, but may fail to track a target owing to various factors, such as the initialization errors and the sudden changes in the target motion. The proposed algorithm can recover the PDAF from failures using an assisting finite impulse response (FIR) filter. The FIR filter operates only when the PDAF cannot track a target properly, and additionally offers state estimate and estimation error covariance to reset the PDAF. The proposed algorithm, the hybrid PDAF/FIR filter (HPFF), combines the PDAF and FIR filter, and hence shows enhanced reliability. Simulations of preceding vehicle tracking using an automotive radar demonstrate the effect and performance of the proposed HPFF.

Hough transform based ego‐velocity estimation in automotive radar system

Hough transform based ego‐velocity estimation in automotive radar system

High-Resolution Delay-Doppler Estimation Using Received Communication Signals for OFDM Radar-Communication System

High-resolution delay-Doppler estimation is an important requirement for automotive radar systems, especially in multi-target scenarios that require better target separation performance. Orthogonal frequency-division multiplexing (OFDM) is a promising candidate waveform for future intelligent transport networks, since it enables the integration of both radar and communication functionalities. Exploring the dual functionality enabled by OFDM, this paper presents a new cooperative method for high-resolution delay-Doppler estimation. The proposed subspace-based method exploits the combination of both the radar and received communication signals to estimate target parameters. The procedure achieves high-resolution delay-Doppler estimation for both uncorrelated, partially correlated and coherent signals, and enables a significant reduction in the required bandwidth when compared to previous approaches which do not exploit the knowledge of the communication signals. Laboratory measurements at 24 GHz and simulation results demonstrate the efficacy of the proposed method for the estimation of multiple targets.

YOLO-Based Simultaneous Target Detection and Classification in Automotive FMCW Radar Systems.

This paper proposes a method to simultaneously detect and classify objects by using a deep learning model, specifically you only look once (YOLO), with pre-processed automotive radar signals. In conventional methods, the detection and classification in automotive radar systems are conducted in two successive stages; however, in the proposed method, the two stages are combined into one. To verify the effectiveness of the proposed method, we applied it to the actual radar data measured using our automotive radar sensor. According to the results, our proposed method can simultaneously detect targets and classify them with over 90% accuracy. In addition, it shows better performance in terms of detection and classification, compared with conventional methods such as density-based spatial clustering of applications with noise or the support vector machine. Moreover, the proposed method especially exhibits better performance when detecting and classifying a vehicle with a long body.

Bi‐static delay‐Doppler reference for cooperative passive vehicle‐to‐X radar applications

Automotive radar systems are indispensable for advanced driver assistance systems and traffic safety. Besides existing monostatic radar techniques, bi-static radar sensing like passive coherent location offers additional options to improve the radar visibility of vulnerable road users. Regarding the testing and evaluation of signal processing algorithms including parameter estimation, it is essential to provide electromagnetically shielded and reproducible measurement conditions, in addition to field tests in real traffic scenarios. This study describes the possibility to emulate relevant performance parameters for bi-static radar scenarios in the frequency range from 1 GHz to 6 GHz in a metal-shielded semi-anechoic chamber. Of special interest are the bi-static angle between transmitter, target, and receiver, and the resulting bi-static Doppler frequencies of a realistic vehicular traffic scenario. According to the concept of cooperative passive coherent location, Doppler scattering measurements are presented and compared to electromagnetic simulations. The authors find promising agreement between measured and ground truth data in the delay-Doppler spectrum.

Omnidirectional retroreflective surface using geodesic polyhedra

In this paper, we present the design and analysis of omnidirectional retroreflective surfaces made of trihedral, corner reflectors to enhance the monostatic radar cross section of otherwise low back-scattering surfaces. The proposed designs can be mounted on helmets and used for various applications such as improving the collision avoidance of automotive radar systems in the presence of bicyclists and motorcyclists or improving collision prediction radars in contact sports or other risky environments, where helmets are used. To address these applications, we analyze the coherent response of the proposed designs.