Neural-network-based terminal sliding-mode control for thrust regulation of a tethered space-tug

Abstract

This paper studies the thrust regulation of the tethered space-tug in order to stabilize the target towed by a exible tether. To compromise between model accuracy and simplicity, a rigid- exible coupling multi-body model is proposed as the full model of the tethered space-tug. More specifically, the tug and the towed target are assumed as rigid bodies, whereas the exible tether is approximated as a series of hinged rods. The rods are assumed extensible but incompressible. Then the equations of motion of the multi-body system are derived based on the recursive dynamics algorithm. The attitude motion of the towed target is stabilized by regulating the thrust on the tug, whereas the tether-tension-caused perturbation to the tug's attitude motion is eliminated by the control torque on the tug. The regulated thrust is achieved by first designing an optimal control trajectory considering the simplified system model with constraints for both state variables and control input. Then the trajectory is tracked using a neural-network-based terminal sliding-mode controller. The radial basis function neural network is used to estimate the unknown nonlinear difference between the simple model and the full model, while the terminal sliding-mode controller ensures the rapid tracking control of the target's attitude motion. Thrust saturation and tether slackness avoidance are also considered. Finally, numerical simulations prove the effectiveness of the proposed controller using the regulated thrust. Without disturbing orbital motion much, the attitude motion of the tug and the target are well stabilized and the tether slackness is avoided.