Modern vehicles have several autonomous and semi-autonomous features, such as adaptive cruise control, lane keeping, adaptive headlights, and automatic emergency braking, ensuring a safe and comfortable driving experience. The vehicles typically rely on different sensors to "see" their surroundings and make decisions accordingly. Among these sensors, radar is particularly significant for its exceptional range and velocity estimation capabilities and plays an essential role in detecting and tracking objects within the vehicle's vicinity.

Since automotive radars operate in the same frequency range, there is a chance that radars operating in close proximity might encounter mutual interference. The interference can degrade the radar's performance and cause false alarms and missed detections, which can be particularly problematic in safety-oriented systems. This research aims to develop signal processing techniques to mitigate the interference effects in frequency-modulated continuous wave (FMCW) radars operating at 77-81 GHz and contribute to making modern radar applications safe and reliable. The interference mitigation methods investigated in this thesis fall into three categories: digital beamforming, time-domain signal reconstruction, and deep learning methods.

The digital beamforming approach utilizes the beam pattern of the receiving antenna array to mitigate interference by placing notches in the beam pattern. It is demonstrated that while this approach is applicable to MIMO radar systems, the notch resolution does not benefit from the extended virtual aperture. An adaptive digital beamforming approach based on the least mean squares (LMS) algorithm is also proposed to suppress interference in the received signal.

The time-domain signal reconstruction approaches aim to reconstruct the parts of the received baseband signal that is corrupted by the interference. It is shown that the signal coherence in the slow-time domain can be utilized to perform signal reconstruction in the slow-time. Moreover, it is shown that by compressing the interference in the time domain using pulse compression, the duration of the interference can be shortened, and an improvement in signal reconstruction performance can be achieved.

Given the complexity of the mutual interference problem, deep learning-based approaches can be instrumental in interference mitigation. This research also investigates the use of deep neural network architectures such as recurrent neural networks, convolutional neural networks, and convolutional autoencoders for signal reconstruction and denoising performance. 

Millimeter-wave (mm-wave) frequency-modu- lated continuous wave (FMCW) radars are increasingly being deployed for scenario perception in various applications. It is expected that the mutual interference between such radars will soon become a significant problem. Therefore, to maintain the reliability of the radar measurements, there must be procedures in place to mitigate this interference. This article proposes a novel interference mitigation technique that utilizes the pulse compression principle for interference compression and mitigation. The interference in the received time-domain signal is compressed using an estimated matched filter. Afterward, the compressed interference is discarded, and the signal is repaired in the pulse-compressed domain using an autoregressive (AR) model. Since the interference spans fewer samples after compression, the signal can be restored more accurately in the compressed domain. Real outdoor measurements show that the interference is effectively suppressed down to the noise floor using the proposed scheme. A signal to interference and noise ratio (SINR) gain of approximately 14 dB was achieved in the experimental data, supporting this study. Moreover, the results indicate that this method is also applicable to situations where multiple interference sources are present.

Automotive radars have become an important part of sensing systems in vehicles and other traffic applications due to their accuracy, compact design, and robustness under severe light and weather conditions. The increased use of radars in various traffic applications has given rise to the problem of mutual interference, which needs to be mitigated. In this paper, we investigate interference mitigation in chirp sequence (CS) automotive radars via signal reconstruction based on autoregressive (AR) models in fast-and slow-time. The interference is mitigated by replacing the disturbed baseband signal samples with samples predicted using the estimated AR models. Measurements from 77 GHz frequency modulated continuous wave (FMCW) static and moving radars are used to evaluate the signal reconstruction performance in terms of the signal-to-interference-plus-noise ratio (SINR), peak side-lobe level (PSLL), and mean squared error (MSE). The results show that the interference is suppressed down to the general noise floor, leading to an improvement in the SINR. Additionally, enhanced side-lobe suppression is achieved via AR signal reconstruction, which is compared to a commonly used inverse-cosine method. Furthermore, the paper notes that the slow-time signal reconstruction can be more beneficial for interference suppression in certain scenarios. CCBY

Automotive radar has emerged as an important sensor for environmental perception in modern vehicles. A rapid increase in the number of radars present in traffic operating at unregulated frequencies has given rise to a mutual interference problem. In order for radar-based systems to function reliably, such interference must be mitigated. In this paper, this problem is addressed with a bidirectional long short-term memory (Bi-LSTM) network as a deep learning approach. Using the Bi-LSTM network, we reconstruct the intermediate frequency (IF) signal and recover samples lost to interference. The proposed signal reconstruction method is evaluated via real measurement data. The proposed Bi-LSTM network provides a better performance than an autoregressive model-based signal reconstruction method. © 2022 European Microwave Association (EuMA).

Mutual interference between automotive radars can make it difficult to detect targets, especially the weaker ones, such as cyclists and pedestrians. In this paper, the interference suppression performance of a Least Mean Squares (LMS) algorithm-based adaptive beamformer is evaluated using measurements from a 77 GHz Frequency Modulated Continuous Wave (FMCW) radar in an outdoor environment. It is shown that the adaptive beamformer increases detection performance and that the interference is suppressed down to the noise floor of the radar in the Range-Doppler domain. In the paper, real baseband sampling and complex-baseband sampling (IQ) radar receivers are compared in the context of interference suppression. The measurements show that IQ receivers are more beneficial in the presence of interference.

This paper addresses the problem of mutual interference between automotive radars. This problem is getting more attention with an increase in the number of radar systems used in traffic. An adaptive digital beamforming technique is presented here which suppresses the interference without the exact knowledge of the interfering signal's Direction of Arrival (DoA). The proposed technique is robust and does not rely on any calibration for the interference cancellation. The adaptive interference suppression method is evaluated using a simulated scenario. Up to about 20-23 dB improvement in the target Signal to Interference and Noise Ratio (SINR) is measured in the simulation and a better detection performance is achieved using the proposed interference suppression technique. © 2018 IEEE.

Vehicle classification has a significant use in traffic surveillance and management. There are many methods proposed to accomplish this task using variety of sensorS. In this paper, a method based on fuzzy c-means (FCM) clustering is introduced that uses dimensions and speed features of each vehicle. This method exploits the distinction in dimensions features and traffic regulations for each class of vehicles by using multiple FCM clusterings and initializing the partition matrices of the respective classifierS. The experimental results demonstrate that the proposed approach is successful in clustering vehicles from different classes with similar appearanceS. In addition, it is fast and efficient for big data analysiS.

Modern vehicles are often equipped with a number of radar sensors that allow the vehicle to be aware of its surroundings for safety and automated driving features. As the number of radars in traffic is growing, mutual interference between radars on different vehicles is becoming unavoidable. Several deep learning methods for mitigating automotive radar interference have been proposed, with the primary differences being data type and network architecture. In this paper, we examine various deep learning-based interference mitigation strategies and investigate the benefits and drawbacks of various model versions. The selected deep learning model variants (convolutional neural networks (CNN) and autoencoders (AE) in time and range-Doppler domain) are trained on a combination of real and simulated radar data with simulated interference. The trained models are then evaluated on a different set of real measurements data with interference from two real sources. The improvement in Signal to Interference and Noise (SINR) in range-Doppler domain methods (~15 dB) over time domain methods (~5 dB) demonstrates excellent denoising capabilities of the range-Doppler domain methods. On the other hand, time-domain approaches are effective at reducing phase errors.