Research On Autonomously Traveling Of Tracked Robot Based On YHFT-DSP

Mobile robot security obstacle avoidance，road recognition and navigation hasalways been a popular research direction in mobile robot field. It has importanttheoretical significance and application prospects. The mult-sensor fusion, whichincludes image messages, provides abundant information for mobile robot securityobstacle avoidance, road recognition and navigation. On this basis, more efficient androbust algorithm can be studied. This topic mainly researches in three aspects: firstly,achieve the multi-ultrasonic sensor obstacle avoidance by the typical method of fuzzingcontrol based on the tracked robot AS-RF. Secondly, implement the road recognition,detection and navigation in the environment of corridor with CCD camera on AS-RF.Thirdly, transfer the project above-mentioned to the platform of high performance floatpoint digital signal processor YHFT-DSP designed by National University of DefenseTechnology to verify its function.Specifically, this paper firstly introduced the development and current researchstatus of mobile robot, analyzed the mobile robotic main research technology and thedifficulties of today. As the increase of sensor types and the complexity of road imageinformation, the development of mobile robot needs real-time, accurate, stable andvalidated mult-sensor control, road image recognition and processing. The embeddedsystem can meet the requirement of the robotic development, adopting the DSP chips asprocessor can guarantee the real-time ability, reduce consumption effectively andimprove the ability of resisting nasty environment.Secondly, this paper introduced the technological process of movement control,sensor distribution and video capture based on AS-RF robot platform. This thesisapplied the fuzzy theory controlling to overcome the incompleteness of sensorinformation collection. Defining the fuzzy sets and fuzzy rules of the obstacle avoidanceto control inputs and outputs, and did experimental verification.Thirdly, this thesis selected the corridor road environment for the research ofAS-RF mobile robot road recognition, detection and processing. By the means of roadimage collection, image enhancement and edge extraction, the robot gained the roadboundary lines in the constraining region of interest and combined corridor roadcharacteristics. This paper designed a method of angle constraint for guiding the AS-RFrobot anti-offset travel and verified algorithm testing the result through experiment.Finally, made a detailed introduction of YHFT-DSP developing board, videocaptureing and processing board. By Reguarding the YHFT-DSP developing board asplatform, this paper has done research on security obstacle avoidance. By regarding theYHFT-DSP video captureing and processing board as platform, this paper has doneresearch about road recognition and processing. At last, the conclusion has verified the theoretical result with experiment.