IC Substrate - Automotive radar and crowded radio spectrum: a potential urban electronic battlefield

With the increasing popularity of automotive radar, the crowded radio frequency spectrum in the urban environment will become an "electronic battlefield." Radars will face combined attacks with unintentional or intentional interference, and designers must implement anti-jamming techniques as in electronic warfare (EW). Automotive radars are usually subject to denial or deceptive interference. Denial jamming blinds the radar of the victim vehicle. This technique will reduce the signal-to-noise ratio, resulting in a decrease in the probability of target detection. On the other hand, deceptive interference will make the victim vehicle’s radar “believe” that there is a false target. The victim vehicle's radar loses the ability to track real targets, so the behavior of the victim vehicle is seriously affected. These interferences may originate from the mutual interference between car radars, or the use of cheap hardware to simply direct strong continuous wave (CW) signals to the victim vehicle radars and deliberate attacks.

Although the current interference avoidance technology may be sufficient to deal with today's situation, with the proliferation of radar sensors, cars will need to use flexible types of mitigation technology, or such technology in combination with interference avoidance methods. Flexible techniques include time-frequency domain signal processing or complex radar waveforms.

Jamming FMCW radar
The radar waveform is one of the key system parameters for judging the performance of the sensor in the presence of interference. Today's automotive radars in the 77 GHz frequency band mainly use FMCW waveforms. In FMCW radar, the CW signal linearly sweeps or chirps on the frequency of the radio frequency band. Figure 1 shows an FMCW chirp sequence (CS) waveform. The frequency difference (fB, beat frequency) of the echo signal is proportional to the distance R to the target, which can be determined by the following relationship:

Among them, fsweep is the change of frequency, and Tchirp is the time of frequency sweep.

In a dense radio frequency environment, interference occurs when the FMCW radar sensor is operating in the same part of the frequency band. A typical example of oncoming car interference is shown in Figure 2a.
Rejected jamming (b) and deceptive jamming (c) of FMCW radar in driving scenario (a)
Rejection interference
Any FMCW-type strong jamming signal falling into the receiver bandwidth will increase the noise floor of the victim radar. Such rejection interference may cause small targets (ie, small radar cross section (RCS)) to disappear due to poor SNR. Refusal interference can also be done deliberately, simply by shooting a strong CW signal at the FMCW radar of the victim vehicle.

Deceptive interference
If the jamming signal scan is synchronized but delayed along with the victim radar, then its effect will be to produce deceptive false targets at a fixed distance (Figure 2c). This technique is very common in electronic warfare jammers. Similar oncoming car radars will become unintentional jammers. However, the probability of time alignment between the victim radar and the jamming radar will be very small. A jammer delay offset that is less than the maximum range delay of the victim radar may look like a real target. For example, the maximum distance of 200m requires the scan alignment error to be less than 1.3 microseconds. However, by installing complex electronic warfare-like equipment on an oncoming car platform, such deceptive attacks can be deliberately carried out.

More generally, deceptive jamming is based on the retransmission of the victim radar signal, but its delay and frequency are systematically changed. This can be incoherent (the jammer in this case is called a transponder) or coherent (the jammer in this case is called a repeater). The repeater receives, changes and retransmits one or more jamming signals, while the transponder transmits a predetermined signal when the jammer detects the target victim radar signal.

Complex attacks based on repeaters usually require digital radio frequency memory (DRFM). DRFM can perform coordinated range delay and Doppler gate drag attacks. Therefore, it will maintain the false target range and Doppler characteristics to deceive the victim radar.

Interference mitigation technology
Basic radar interference mitigation techniques mainly rely on methods of avoiding interference. The goal is to reduce the possibility of space, time and frequency overlap, for example:

* Space: The use of a narrower electronic scanning beam can reduce the risk of interference. The typical field of view of the long-range car cruise control (ACC) radar is ±8 degrees. Nevertheless, strong interfering signals can still cause effective interference through antenna side lobes.

* Time: FMCW chirp slope parameters are randomly generated to avoid periodic interference.

* Spectrum: Randomly generate FMCW chirp start and stop frequencies to reduce the probability of overlap and interference.

The basic method of randomization will avoid accidental synchronization with other radars, but may not be so useful in dense RF environments. More and more radar sensors require more complex and flexible technology to mitigate interference.

Detect and repair
Another way to avoid interference is to use signal processing algorithms to repair the received waveform. Time-frequency domain technology can effectively deal with denial-type jamming attacks. In the oncoming car scene (Figure 2), the jammer scans all frequency bins for a very short time. This fast time-varying signal appears as a raised noise floor in the conventional FFT domain. The time-frequency domain signal processing technology transfers the signal to another domain. Compared with the FFT domain, it is easier to filter out interference in this domain.

For time-varying signals, short-time Fourier transform (STFT) can provide more information than regular FFT. STFT-based technology can be used to eliminate narrowband interference (see Figure 3). STFT basically moves a window through the signal and obtains the FFT of the window interval. The signal is filtered in the frequency domain to remove interference components, and then converted back to the time domain. Figure 4 shows a typical FMCW interference situation with overlapping radio frequency chirp sequences, and the resulting IF beat signal in the STFT domain. The IF domain is shown on the right, which is the final result of mixing the radar (blue) and interference (orange) signals. The horizontal line indicates the target, and the V-shaped vertical line indicates the presence of interference signals. Interference FMCW in the same or opposite direction, or even slow chirp similar to CW, has similar effects on the IF signal. In all these interference situations, the fast-moving V-shaped IF signal will increase the noise floor in the regular FFT domain.

Amplitude-based masking can be used to filter out interfering signals in the STFT domain. Of course, the premise is that the front end and quantization part of the victim radar have sufficient dynamic range to simultaneously linearly process the stronger interference signal and the smaller expected target signal. Figure 5a shows a strong interference signal, and Figure 5b shows the processed STFT. In the case of strong interference, as shown in Figure 5a, multiple real targets are not visible. In Figure 5b, the V-shaped interference signal is eliminated; when transferred back to the time domain, the low SNR target is now identifiable.
In the rejection-type interference situation, STFT-based interference mitigation technology can be used to deal with strong interference. For deceptive jamming attacks, STFT alone cannot verify whether the return signal is true or false.

Encrypted radio frequency
The basic countermeasure to reduce the impact of repeater deceptive jamming attacks is to use low probability intercept (LPI) radar waveforms. The purpose of LPI radar is to spread the radiated energy over a wide frequency spectrum to avoid detection, usually using quasi-random scanning, modulation or frequency hopping sequence. FMCW is an LPI waveform. If phase encoding or encryption is introduced into frequency chirp, it can further reduce the probability of DRFM intercepting automotive radar signals. The unique encrypted radio frequency characteristics of each radar sensor can verify the authenticity of the returned signal.

Two of the same radars (installed on different cars) have frequency offsets and delays between them, creating a false target in the victim radar. The jamming radar and the victim radar are aligned in time (same chirp slope and shorter offset). In this case, the phase-coded FMCW radar can provide high anti-jamming capabilities. The use of orthogonal codes also makes MIMO radar operations feasible, thereby supporting simultaneous transmission of multiple waveforms.