Dynamic Soaring Trajectory Optimization Considering the Path Following Performance

Abstract

Dynamic soaring is a technique employed by some kinds of birds to extract energy by intelligently flying in shear wind. The combination of trajectory optimization and path following is currently the most used duo in real-world application of dynamic soaring. However, the separate design of the trajectory and the controller inherently introduces a fair degree of conservativeness, compromising optimality. In this study, we propose a dynamic soaring trajectory optimization algorithm, which additionally considers the closed-loop stabilization of the path following error dynamics through gain scheduling. In other words, the stability of the path-deviation error controller is taken into account through the Routh criterion already in the form of constraints when designing the trajectory in the optimization problem. Consequently, the optimized dynamic soaring trajectory as well as the feedback gains are obtained simultaneously by solving the trajectory optimization problem. The proposed algorithm balances optimality and stability and thereby reduces conservativeness. Eventually, simulations demonstrate the effectiveness of the algorithm.