Flight control method based on learning

Flight control method based on learning

The characteristic of learning-based flight control method is that it does not need to know the dynamic model of the aircraft, but only needs some flight tests and flight data. Among them, the most popular research methods are fuzzy control method, method based on human learning and neural network method.

1. Fuzzy logic Fuzzy control is one of the methods to solve the uncertainty of the model, and it can control the UAV when the model is unknown.

2. Human-based learning In order to find a better control method for small unmanned aerial vehicles (UAVs), researchers at MIT in the United States collected data from aerobatics aircraft participating in military exercises, and analyzed pilots' operation of aircrafts in different situations, so as to better understand the input sequence and feedback mechanism of UAVs. This method has been applied to the autonomous flight of small UAV.

3. The classical PID control of Neural networks is simple in structure, easy to use and realize, but when the controlled object has complex nonlinear characteristics and it is difficult to establish an accurate mathematical model, it is often difficult to achieve satisfactory control results. Neural network adaptive control technology can effectively control a variety of uncertain and difficult-to-describe nonlinear complex processes, improve the robustness and fault tolerance of the control system, and the control parameters have adaptive and self-learning ability.