Inter‐radar interference analysis of FMCW radars with different chirp rates

Abstract

Millimetre-wave frequency-modulated continuous wave (FMCW) radar is expected to be widely deployed for advanced driver assistance systems (ADAS) and self-driving cars. Considering that automotive radars will be deployed in a huge number of cars in future, inter-radar interference will be a significant problem since inter-radar interference can cause miss-detection and/or false detection of the target. As millimetre-wave automotive radar is not standardised at this moment, it is required to evaluate inter-radar interference among FMCW radars with various chirp-rates and chirp durations. This study describes inter-radar interference analysis in FMCW radars with different chirp rates and simulation results of inter-radar interference to confirm the validity of the interference analysis. 11Introduction In recent years, aiming at safer and more efficient transportation, research and development of ADAS and automated driving technologies is widely carried out all over the world [1, 2]. In ADAS and automated driving car, car driving is partly performed by the computer on-board the car using sensor information and other information such as location information and MAP data. Therefore, ADAS and automated driving car need various types of sensors such as optical cameras, laser imaging detection and ranging (LiDAR), radar, ultra-sonic sonar to recognise the surroundings of a car. As each sensor has advantages and disadvantages, multiple types of sensors are used in combination. Regarding radar, millimetre wave radar can use wide spectrum of 4 GHz in 79 GHz band, thus distance measurement accuracy of 10 cm can be achieved. One of the promising radar for automotive applications is millimetre-wave FMCW radar because it achieves fairly high distance resolution of about 10 cm, and it can simultaneously measure the relative distance and relative speed of the target. Furthermore, direction of the target can be measured by using multiple-input multiple-output (MIMO) configuration and it is relatively inexpensive compared with LiDAR. Therefore, it is expected to use several millimetre-wave FMCW radar in an automated driving car. Considering that automotive FMCW radars will be densely deployed in future, inter-radar interference can be a serious problem since it causes miss-detection of the target and/or false detection of the so-called ghost target. It is well-known that there are two types of radar interference, i.e. narrowband interference and wideband interference [3]. Narrowband interference occurs when the interference radar use the same chirp rate and chirp direction as the observation radar, and the observation radar detects a target that does not exist as a ghost target. On the other hand, wideband interference occurs when the interference radar has different chirp rate from the observation radar, resulting in an impulse-like interference signal in time domain. Fourier transform of impulse-like interference signal causes noise floor increase in frequency domain. Thus, SNR degradation due to wideband interference causes miss-detection of the target. Regarding wideband interference, various inter-radar interference mitigation methods have been proposed, e.g. frequency hopping random chirp (FHRC)-FMCW radar to avoid narrowband interference [4] and weighted envelope normalization (WEN) algorithm to reduce wideband interference [5]. The authors also proposed the interference detection and suppression method in the time domain to mitigate wideband interference [6]. As wideband interference mitigation methods are based on interference suppression in time domain, SNR improvement by interference reduction depends on received level and design parameters such as chirp rate of interference radar. As there is no standard of millimetre-wave FMCW radar so far, a large number of FMCW radar with different design parameters can be used simultaneously. Therefore, it is necessary to evaluate SNR degradation due to inter-radar interference among a large number of FMCW radars with various design parameters. This paper describes inter-radar interference analysis of FMCW radars with various design parameters to estimate SNR degradation caused by wideband interference. The results of wideband interference analysis are confirmed by computer simulations. This paper reveals that inter-radar interference becomes significant when the chirp rate of interference radar is similar to that of interference radar. The inter-radar interference analysis implies that standardisation of FMCW radars is desirable to mitigate wideband interference. 22Interference in fast chirp FMCW radars 2.1 Fast chirp FMCW radar Fig. 1 shows a block diagram of millimetre-wave FMCW radar. Frequency modulated linear chirp signal is transmitted, and reflected signal from the target is received with delay time. It is mixed down by multiplying the transmitted radar signal to extract beat signals, whose frequency is in proportion to the distance of the target. To detect the frequency of beat signals, AD conversion and fast Fourier transform (FFT) processing are performed and the peak of the frequency, i.e. distance of the target, is detected as the peak in frequency spectrum. There are two kinds of modulation waveform of FMCW radar, i.e. triangular wave modulation and saw-tooth wave modulation. The triangular wave modulation uses both up chirp and down chirp to calculate the relative speed and relative distance of the target, however, peak matching of up-chirp and down-chirp becomes difficult when there are many targets to be detected. On the other hand, the saw-tooth wave modulation uses either up-chirp or down-chirp, and the fast chirp modulation with short chirp period is considered to be promising for multi-target detection [7]. In the fast chirp radar, each peak is tracked target by target and velocity of