A Model-Free Control Algorithm Based On The Sliding Mode Control Method With Applications to Unmanned Aircraft Systems

Abstract

Control methods require the use of a system model for the design and tuning of the controllers in meeting and/or exceeding the control system performance objectives. However, system models contain errors and uncertainties that also may be complex to develop and to generalize for a large class of systems such as those for unmanned aircraft systems. In particular, the sliding control method is a superior robust nonlinear control approach due to the direct handling of nonlinearities and uncertainties that can be used in tracking problems for unmanned aircraft system. However, the derivation of the sliding mode control law is tedious since a unique and distinct control law needs to be derived for every individual system and cannot be applied to general systems that may encompass all classifications of unmanned aircraft systems. In this paper, a model-free control algorithm based on the sliding mode control method is developed and generalized for all classes of unmanned aircraft systems used in robust tracking control applications. The model-free control algorithm is derived with knowledge of the system’s order, state measurements, and control input gain matrix shape and bounds and is not dependent on a mathematical system model. The derived control law is tested using a high-fidelity simulation of a quadrotortype unmanned aircraft system and the results are compared to a traditional PID controller for tracking performance and power consumption. Realistic type hardware inputs from joysticks and IMUs was simulated for the analysis. Finally, the model-free control algorithm was implemented on a quadrotor-type unmanned aircraft system testbed used in real flight experimental testing. The experimental tracking performance and power consumption was analysed and compared to a traditional PID-type controller.