Adaptive Waveform Design Based On Information Theory

The transmit waveform of conventional radar is normally fixed, which limits the performance in complex battlefield. With the recent hardware development of high speed signal processer and power amplifier, the agility of transmitting arbitrary waveform is possible. The adaptive waveform design which allows the radar to actively adjust its paramenters to the changing target and environment in complex electromagnetic battlefield has significant application in military.In this thesis we focus on the highly intelligent trend of radar system. Information theory based radar waveform optimization problem is investigated to improve the radar performance. Besides, the radar waveform optimization methodology is applied to the optimization of jamming signals, so that further valuable techniques and conclusions are obtained.Resolution is a critical issue in radar systems and is closely related to the target detection and tracking performance. Firstly, the traditional ambiguity function based radar nominal resolution is reviewed, followed by the analyses of several practical resolution which takes into account the measurement model and measurement and estimation error. Based on the detection theory and information resolution concept, the practical resolution in the sense of Kullback-Leibler divergence is proposed, which incorporates the ambiguity function property of the waveform as well as the measurement model and measurement moise. Thus, it is also an extension of the Woodward norminal resolution. Secondly, the radar waveform parameter optimization based on the Kullback-Leibler divergence criterion is proposed. The optimal waveform can improve the practical resolution of two closely located point targets. Finally, this approach is extended to the bistatic radar case, where the optimal waveform parameter and the bistatic baseline range are jointly designed. Radar practical resolution is enhanced using the energy constraint waveform optimization in this chapter. The research constructs the premise and foundation of radar waveform design for target detection, tracking and identification.Target identification is an important task as well. Firstly, Chernoff divergence between the probability density function of the two target measurements is analyized which affects the error probability of two point tagets classification. By maximizing the Chernoff divergence, the optimal waveform parameters are adaptively changed so that the classification error is reduced in a great deal. Secondly, the multiple extended targets identification problem is investigated. This problem can be regarded as a multiple hypothesis testing problem, and sequential likelihood ratio test is applied to solve this problem. Several radar illuminations are adopted and the final decision is made based on likelihood ratio test and the terminal condition. At each illumination, the Bayesian method is used to update the current probability of each hypothesis. Therefore, the optimal waveform is adaptively changed according to the recent information. By optimally changing the waveform, the average illumination number needed for target identification is decreased. Finally, Multiple-Input Multiple-Output(MIMO) radar waveform optimization to improve target identification performance is studied, and optimal power allocation strageties are obtained for MIMO radar. The concousion we find extendes the Single-Input Single-Output(SISO) radar optimal result for target identification. Through the analyses in Chapter 2 and Chapter 3, optimal radar waveform for two or multiple target resolution and identification problem are obtained, which complement and extend the existing methods in literature.Finally, the representative radar waveform optimization methods and conclusions are applied to the problem of optimal jamming technique for smart jammer. Smart jammer is an active device which can adaptively change is transmitted jamming signal according to the intercepted radar and target information, so that the jamming performance is enhanced. The basic principle of optimal jamming technique greatly resembles that of the radar waveform optimization. Therefore, the information theory based radar waveform design can be used in optimal jamming techniques, which extendes the application of radar waveform optimization, and further founds the basis of integreated radar and jammer research and the research of radar and jammer waveform design when they are coexistent. In this thesis, the optimal jamming technique based on signal-to-jamming-plus-noise ratio and mutual information is proposed respectively. Optimal jamming power allocation strategies are obtained for difference jamming purpose. In the previou jamming techniques, the prior information is assumed to be perfectly known, which is impossible in practice owing to the fact that the information are estimated and errors will exist. Therefore, robust jamming techniques are investigated which maximize the worst jamming performance under the uncertain information scenario. The research in this chapeter makes us a more profound understanding of the information theory based radar waveform optimization. The joint design of radar and jammer signals when the radar and jammer are affected by each other will be studied in future accoding to the methods and results in this thesis.