Radar to radar interference for 77GHz automotive radar

Abstract

Automotive radar is a key element in Advanced Driver Assistance Systems (ADAS). With the growth of Automotive industry, there is a high demand for the sensors used in assisting systems. As the number of sensors increases, probability that these systems being in close proximity will also increase. This will lead to situations where in multiple radar sensors will be operating in close proximity, this might lead to a sub-optimal performance of our radar system. Currently FMCW radar systems are most prevalent in the automotive radar market. For FMCW systems, interference mitigation techniques exist in time domain, frequency domain, polarization domain and etc. Most of these techniques insiston detecting and identifying the interference before mitigating it. In this thesis, we consider a FMCW radar system and first develop a MATLAB model for the three most important interference scenarios namely Continuous Wave(CW), FrequencyModulatedContinuousWave (FMCW) and PhaseModulatedContinuousWave (PMCW). We propose a signal model where interference can be localized over the beat signal, then we systematically study how interference can be detected. As a result, we will detect the interference even if the power level of the interferer is lower than the power level of the received reflected signal. Post detection of interference,we suggest a technique to identify the interference by estimating the slope from the detected interference samples. Starting with a simple existing mitigation technique, we look at how to mitigate the interference and suggest enhancements that can be done for these techniques post detection and identification. Considering a worst case scenario of interference being completely in band to the transmitted signal, we propose a novel avoidance technique which will also predict the bandwidth of the interferer. As a result we will be able to shift the center frequency of the transmitter to avoid the interferer. Finally we propose another novel time domain mitigation technique where in without detecting or identifying the interferer, we will mitigate the interference and compare the gain of Signal to Interference plus Noise Ratio (SINR) achieved by using this technique.