The Controller Research Based On The Behavior-robot

Since the birth of the robot, people have been thinking of that one day, we can befree to communicate with the robot. With the development of the science andtechnology, robot by the initial repetitive operators have gradually transformed intothe certain intelligent role. But due to the limitation of conditions, have not been ableto have a major breakthrough. Since the1980s, Brooks et al introduce the Behaviorinto the robotics field; robot development has entered a new period of development.During this period, many new behavior methods have been proposed, and validatedby experiment related, and achieved good results of scientific research and the use ofeffects. Behavioral robotics research has become a new hotspot.The article firstly introduce the development of robot, and then analyzes thebehavior of the robot controller development and research status at home and abroad,and then the introduced and analysis the behavior of the robot control methods, pointout that the behavior of robotics research focus on the planning method for robotbehavior and behavior selection mechanism. Existing robot behavior selectionmechanism can do very good to make the robot obstacle avoidance and dynamic&superficial state environment etc, and bas been done many powerful experiment andpractical application. But also have their shortcomings, namely, there’s no method tolet the robot learning-itself when it is moving.Above all, the article has carried out based on the behavior of the robot controllerof exploration and research; put forward a kind of robot online learning-itself method.The behavior of the robot programming method respectively to compare and merge,to extract a series of basic behavior pattern, used to secure and reliable behavior robotmotion control. The basic behavior by changing the behavior or merge can to create anew task to complete more complex action behavior. At the same time, try to contrastto the commonly used several kinds of behavior selection mechanism, analyzes thecharacteristics of their own, and point out their applicable scope, behavior of robotonline self-learning method is put forward. Through a barrier in a simple environmentsimulation, verify the feasibility and effectiveness of the proposed method. Design the robot controller experiment equipment by myself, which was basedon easy to exam the effect of the arithmetic, according to the need of the robotbehavior-plan and behavior-choice method’s checking. The robot may “feel” thesurroundings by the ultrasonic sensor. At last, do the experiment of the robotlearning-itself method in real-time and the total design of the robot, analyze the result,get the expect efficiency.