Research On Trace Tracking And Branch Identification By Multi-Information Fusion

Trace following and branch identification technology is the most important technology inmany fields, such as self-direction of vehicle, unmanned drive, trace robot and trace finder, etc.The key of trace following and branch identification technology is obtaining accurate andreliable trace information.Based on the analysis of current research of trace tracking, a method of researching on tracetracking and branch identification by multi-information fusion is put forward. Trace trackingsystem with microprocessor STC12C5410AD as the core is designed, eight infrared sensorsare adopted to detect trace information, and trace information is sent to SCM after amplify,filter and A/D transform. In the information process, BP neural network model of trackingsystem is built, and direction and degree of tracking device deviating from trace are obtainedby BP fusion algorithm, then, tracking of tracking device is controlled according to controlalgorithm of motor turning speed and turning direction.In the design, trace branch identification methods are studied based on trace shapes. Forblind zone of trace, linear interpolation and quadratic interpolation are adopted to get trackingposition information of tracking device, and concept of trace genetic information is putforward. Trend of trace can be estimated by regression algorithm, and control decision can bemade combing detected information, thus, tracking is adaptive to outside environment.Meanwhile, self-study technology of trace information is also put forward, tracking device cantrack trace repeatedly through self-study of special trace, that is to say, tracking device hastrace-memory function.The system is proved to be feasible after some tests and MATLAB simulation, and hightrace tracking accuracy can be achieved.