Small unmanned helicopter fixed-height control method based on reinforcement learning

Abstract

The invention relates to an intelligent height control method for an indoor small unmanned helicopter, and provides a continuous control method based on offline reinforcement learning to realize fixed-height flight of the small unmanned helicopter in an indoor positioning environment under the condition of external disturbance. Therefore, the technical scheme adopted by the invention is as follows: the small unmanned helicopter fixed-height control method based on reinforcement learning comprises the steps that firstly, a helicopter model and a return function based on a Markov sequence are constructed, then a random approximation method is adopted for training iteration to obtain optimized controller parameters, and finally a trained controller is brought into a real helicopter system tobe controlled. The method is mainly applied to indoor small-sized unmanned aerial vehicle design and manufacturing occasions.

Description

technical field [0001] The invention relates to an intelligent control method for height control of an indoor small unmanned helicopter, in particular to an enhanced learning control method based on off-line flight data. Background technique [0002] When a small unmanned helicopter performs such as hovering, low-speed forward flight and cruising flight, it needs to have the ability to stabilize its own altitude. The fixed altitude control is also the basis for the stable flight of the helicopter and many other complex controls. However, the design of height controllers also faces many serious problems. First, the altitude maintenance of the aircraft cannot achieve the goal through the control of the pitch angle. When the aircraft is subjected to a constant disturbance moment on the longitudinal channel, the flight velocity vector will gradually deviate from the original direction, thus inducing altitude drift. Second, in the dynamic process when the pitch angle tends to b...

AI Technical Summary

Problems solved by technology

Second, in the dynamic process when the pitch angle tends to be stable, when the average value of the change in the inclination angle of the navigation track is not zero, it will also cause a change in the flight altitude

Third, when the helicopter is at a fixed height, since the main rotor itself can generate a large air flow, it will cause a large drift disturbance to the stable state of the fuselage

Fifth, from the analysis of control operability, although the flight height can be controlled by controlling the elevator or the thrust of the engine, the inertia of controlling the flight height by controlling the thrust is very large and the response is slow

Although this method can achieve stable control of the system, it has no constraints on the speed and accuracy of error convergence, and the control effect is relatively general.

Images