TS Fuzzy Model Based On Fuzzy Clustering And Its Application To The Control Of Pan-Tilt In Unmanned Aerial Vehicle

As a new technique of remote sensing, low-altitude UAV (Unmanned Aerial Vehicle) remote sensing systems have been widely used in environmental monitoring, natural disaster relief, etc. In most cases, data acquisition device such as camera is installed in onboard pan-tilt, namely, the quality of acquired data is dramatically depend on the stability of pan-tilt. As a result, how to device a stable controller for pan-tilt becomes one of the most challenging topics in low-altitude remote sensing area. Motivated on this, in this paper, we focused on the design of TS fuzzy control system and pan-tilt control strategy. The main contributions of this dissertation are summarized as follows:1. A new TS fuzzy model identification algorithm based on fuzzy clustering is presented. Firstly, the proposed approach uses modified mountain clustering algorithm to identify the number of fuzzy rules. Secondly, the optimal input-output space fuzzy partition matrix and initial values of system parameters are obtained by the FCRM algorithm. Finally, the system parameters are fine adjusted by the well-known gradient descent algorithm. Simulation results have shown that the proposed approach can be an efficient tool for devising fuzzy identification systems. It is worth noting that the proposed method doesn’t need a predefined number of fuzzy rules. On the contrary, we use optimization algorithm to determine the number and thus the accuracy of identification has been improved.2. According to the features of pan-tilt control, a new hybrid fuzzy controller is proposed which combines TS fuzzy controller with PID controller. The basic idea is to use fuzzy TS system to tune the parameters of PID controller. Simulation results show that our hybrid is faster than traditional approaches, especially when the parameters and the structure of pan-tilt have been changed.