物理学报2018，Vol.67Issue(5)：31-40,10.DOI:10.7498/aps.67.20171185
基于连续时间广义预测校正的水下非线性追踪博弈控制
Differential game guidance of underwater nonlinear tracking control based on continuous time generalized predictive correction
摘要
Abstract
In order to overcome the difficulty in real-time effectively acquiring the target parameters of differential game guidance in a complex underwater environment, the differential game guidance of underwater nonlinear tracking control based on continuous time generalised predictive correction is proposed. Since the target parameter and the detection precision are seriously affected by the acoustic homing device detection period, noise, and interference, it is easy to lose or misjudge the target signal. Hence a combination of the dynamic tracking game model for differential games and the acoustic homing detection method of underwater tracking is used for making the on-line prediction and compensation correction to the deviation tendency of target manoeuvres deviating from the self-guided sound zero axis. This is carried out by using a continuous time generalized predictive control algorithm, according to the discrepancy between the predicted advance angle and the expected value. The manoeuvring target can then be located in the maximum capture probability sector of the tracker device in real time. In order to solve the estimation difficulty problem of the remaining time of the dynamic differential game antagonism, and improve the response speed and the control precision of the system, the zero-efficiency control parameter and the predictive control algorithm are introduced to optimize the differential game. In this way, the infinite time domain differential game can be transformed into a multiple-time domain differential game with feedback correction. Through the complementing advantages of dynamic programming and predictive optimization, the real-time compensation and correction to the interceptor differential game guidance is realised, and the disadvantages of the differential game in the process constraints and stochastic disturbance are overcome. In order to adjust the favourable advance angle of the self-guided detection rapidly, the learning prediction function of rolling optimization feedback correction is adopted. The initial moment of the differential response is pushed forward along with the entire forecast period by rolling optimization. To verify the validity of the algorithm, this is applied to the underwater nonlinear tracking game, and the guidance performance is compared with the differential game guidance and the integrated control algorithm of differential game and discrete predictive control. The results show that this can achieve the optimum control of the high precision underwater manoeuvring target on-line tracking and prediction correction with the detection mode limited in uncertain disturbances, because this is flexible in the choosing of sampling time and does not need control weighting for non-minimum phase system. This can also solve the problem of the initial bias and random disturbance taking into account the control constraints and interference suppression performance, and can improve the robustness to environmental interference.关键词
连续时间广义预测控制/追踪/博弈控制/补偿校正Key words
continuous time generalized predictive control/tracking/differential game guidance/compensated correction引用本文复制引用
程瑞锋,刘卫东,高立娥,康智强..基于连续时间广义预测校正的水下非线性追踪博弈控制[J].物理学报,2018,67(5):31-40,10.基金项目
国家自然科学基金(批准号:61473224)和国家重点研发计划(批准号:2016YFC0301700)资助的课题.Project supported by the National Natural Science Foundation of China (Grant No. 61473224) and the National Research and Development Plan (Grant No. 2016YFC0301700). （批准号:61473224）
