Let's share CommRad: Effect of radar interference on an uncoded data communication system

Spectrum sharing between radar and communications systems is currently under investigation because of the high demand for new wireless services and shortage of available bandwidth. To effectively design their coexistence, it is crucial to understand how current unaltered radar and communication systems would affect one another. This paper investigates the effect of radar interference on an uncoded data communication system where the optimal Maximum-A-Posteriori decoder is used and where the bandwidth of the radar system is much larger than the one of the communication system. Conclusions depend on how the radar interference power, measured by the Interference-to-Noise ratio (INR), compares with the intended signal power, measured by the Signal-to-Noise ratio (SNR). For the case of real-valued modulation schemes, three regimes emerge: (a) Treat interference as Gaussian noise: when INR<;SNR it is optimal to use the threshold decoder for Gaussian noise only. The probability of error increases with INR. (b) Interference cancellation: when INR>>SNR the optimal receiver estimates the radar interference and subtracts it from the received signal; in the process of canceling interference, part of the useful signal is also cancelled, which reduces the effective SNR at the receiver. The probability of error exhibits an irreducible error floor, which can be exactly characterized and behaves like a narrow-band fading channel with multiplicative fading that is perfectly known at the receiver. (c) When INR≈SNR, the probability of error attains its maximum value, thus indicating that there is a worst operating INR for any given SNR.