Neural-Adaptive Constrained Flight Control for Air–Ground Recovery Under Terrain Obstacles

This article contrives a neural-adaptive constrained controller of the cable towed air–ground recovery system subject to terrain obstacles, unmeasurable tensions, trailing vortex, wind gust, and actuator saturation. In modeling, vehicle's nominal 6 DOF affine nonlinear dynamics system's finite links-joints are formulated. To achieve accurate under an asymmetric barrier Lyapunov function-based flight vehicle is proposed, by transforming obstacles into time-varying constraints on trajectory. Then, approximate lumped unknown caused tensions airflows, several echo state network (ESN) approximators established for velocity attitude subsystems. By using approximation errors-based neural weights learning strategy minimal parameter technique, these ESNs possess better transient behaviors lower online computational burden. Furthermore, saturation automatically monitored released, incorporating specially designed auxiliary compensating angular rate control law compensation. The stability closed-loop analyzed. Finally, numerical simulations two scenarios performed demonstrate performance proposed controller.