Research And Application On Fuzzy Pattern Recognition In Driving Simulator

Firstly, this dissertation provides a review on the pattern recognition techniques for image, sound, wave-shape, symbol, and text etc. According to the features of recognizing a simulated operating process , the recognition methods and features of these objects characterized as an one-dimension data or a symbol sequence have been compared in detail, and the adaptability of the recognition methods to action sequence recognition has been analyzed.According to the features that all basic actions in operation are a time-relationship sequence, an one-dimension symbol sequence is used to describe a simulated operating process. By giving corresponding time and sequence attribute to every basic action, the pattern recognition problem in operating process is conversed to the problems of both the adscription of a single acting element to various operation patterns and how to divide the action sequence into a set of subsequences waiting for recognition.In actual recognition process, fuzzy set is used to describe the feature of each operating pattern in simulating operation, a method judging adscription-possibility of any action in infinite long sequence to each operating pattern class according to its sequence attribute and time attribute in each operating pattern has been proposed. It can assign any action in infinite long sequence to a corresponding pattern class and divide an infinite long sequence relating to real operating process into several subsequences waiting for recognition. Finally, pattern of the action sequence segment can be confirmed by comparing possibility scale of subsequences waiting for recognition.A recognition simulation for 11 operation patterns in automobile driving simulating process is made using the method mentioned above, the result of the simulation shows that the operation process done by operator can be recognized correctly in the cases of there are several different operating error. The recognition effects are satisfied.This dissertation introduces an evaluating model of simulating operation quality based on the multi-layer fuzzy evaluation method, and improves the comparison method of confirming significance factor of each evaluation elements. Combining this improved comparison method with improved AHP index scale method, the consistency of judgment matrix in process of confirming weight of each elements is improved.Finally, this dissertation introduces status of research and applications on driving simulator and advantages of using driving simulator in automobile driving training. The active simulator with a self-help function developed by author has also been described in detail.