This paper introduces an optimal autopilot design technique based on the constrained optimization. The tracking performance is formed analytically as the design objective. The open-loop crossover frequency and the maximum demand of the actuator fin-deflection rate are introduced as analytical inequality constraints. | Journal of Automation and Control Engineering Vol. 5, No. 1, June 2017 Optimal Design for Three-loop Autopilot Using Multi-Constraint Optimization Mohamed A. Abd-Elatif* and Qian Longjun School of Automation, University of Science and Technology, Nanjing 210094, China Email: Mdyosf2010@, qlongjun@ value [1,4]. Equally important, the physical limits of the control surface actuators should be considered; otherwise, the performance of the closed-loop system may be significantly degraded or may even become unstable [5]. Basically, classical design approach achieves the desired performance and provides a sufficient stability and robustness [4,6]. This method could be consider as a closed-loop pole assignment technique where the desired poles in the closed-loop characteristic are adjusted to satisfy the design performance in both time and frequency domains. However, this technique is valid for a limited set of flight conditions and system size. Besides, the classical design approach does not guarantee an ideal autopilot as it mostly depends upon the designer manual tuning. On the other hand, LQR is widely recognized as a main design tool in modem control theory. The idea of combined the optimal with classical design aspects shall provide a system that satisfies design constraints requirements and at the same time minimize a performance criterion. For example, optimal autopilot design is introduced based on weightings adjustment for a minimum error between desired and the actual open loop crossover frequency [7-9]. However, this method relies on initial guessing of the weights which might need to be carried out and repeated to adjust the required initial performance. Moreover, this scheme will not essentially guarantee an optimal autopilot as it is possible to get the same crossover frequency for different gain designs. For more realistic applications, the actuator fin deflection rate and the crossover frequency value are considered through a .
