Mitigation of nonlinear receiver effects in modern radar: advanced signal processing techniques
This thesis presents a study into nonlinearities in the radar receiver and investigates advanced digital signal processing (DSP) techniques capable of mitigating the resultant deleterious effects. The need for these mitigation techniques has become more prevalent as the use of commercial radar sensors has increased rapidly over the last decade. While advancements in low-cost radio frequency (RF) technologies have made mass-produced radar systems more feasible, they also pose a significant risk to the functionality of the sensor. One of the major compromises when employing low-cost commercial off-theshelf (COTS) components in the radar receiver is system linearity. This linearity trade-off leaves the radar susceptible to interfering signals as the RF receiver can now be driven into the weakly nonlinear regime. Radars are not designed to operate in the nonlinear regime as distortion is observed in the radar output if they do. If radars are to maintain operational performance in an RF environment that is becoming increasingly crowded, novel techniques that allow the sensor to operate in the nonlinear regime must be developed. Advanced DSP techniques offer a low-cost low-impact solution to the nonlinear receiver problem in modern radar. While there is very little work published on this topic in the radar literature, inspiration can be taken from the related field of communications where techniques have been successfully employed. It is clear from the communications literature that for any mitigation algorithm to be successful, the mechanisms driving the nonlinear distortion in the receiver must be understood in great detail. Therefore, a behavioural modelling technique capable of capturing both the nonlinear amplitude and phase effects in the radar receiver is presented before any mitigation techniques are studied. Two distinct groups of mitigation algorithms are then developed specifically for radar systems with their performance tested in the medium pulse repetition frequency (MPRF) mode of operation. The first of these is the look-up table (LUT) approach which has the benefit of being mode independent and computationally inexpensive to implement. The limitations of this communications-based technique are discussed with particular emphasis placed on its performance against receiver nonlinearities that exhibit complex nonlinear memory effects. The second group of mitigation algorithms to be developed is the forward modelling technique. While this novel technique is both mode dependent and computationally intensive to implement, it has a unique formalisation that allows it to be extended to include nonlinear memory effects in a well-defined manner. The performance of this forward modelling technique is analysed and discussed in detail. It was shown in this study that nonlinearities generated in the radar receiver can be successfully mitigated using advanced DSP techniques. For this to be the case however, the behaviour of the RF receiver must be characterised to a high degree of accuracy both in the linear and weakly nonlinear regimes. In the case where nonlinear memory effects are significant in the radar receiver, it was shown that memoryless mitigation techniques can become decorrelated drastically reducing their effectiveness. Importantly however, it was demonstrated that the LUT and forward modelling techniques can both be extended to compensate for complex nonlinear memory effects generated in the RF receiver. It was also found that the forward modelling technique dealt with the nonlinear memory effects in a far more robust manner than the LUT approach leading to a superior mitigation performance in the memory rich case.