An Endocrine System Inspired Behavior Controlling Algorithm For Robots System

With the development of robot, the demand of robots, especially intelligent and multifunctional robots, has been increasingly high. The moving robot system has to interact with the environment during the working process. When different items in the environment such as different obstacles,different goals and the energy of robot,influence the robot, it has to work with different behaviors accordingly. One of the key and most challenging issues when developing autonomous systems which are able to adapt tounforeseen situations and disruptions is how to coordinateconflicting behaviors and to operate and remain adaptedwithin a viability zone.The nervous system, the immunologic system and the endocrine system shows great ability to deal with different kinds of information. It means a lot in the field of artificial intelligence. Regulatory mechanisms inspired on endocrine system show great adaptiveness and stability. And it been successfully applied in many researches. To solve the behaviors controlling problem during different kinds of environment, this paper propose a set of system inspired on endocrine system with artificial hormones, behavior controller and hormone receptors. With the use of the system, we can make the robot system work efficiently and keep the system stable. The main contributions of the thesis are as follows(1)Solve the problem of single behavior with single robot in obstacle avoidance. Introduce the main idea and the merits and drawbacks of artificial potential field method in the beginning of this chapter. Based on its drawbacks, we propose an artificial endocrine system with one kind of hormone, one kind of receptor and two kinds of behaviors controller. Then we improve the conventional algorithm in 3 aspects. The simulation experiment verifies the accuracy of this algorithm. After the contrast experiment with conventional algorithm, it shows that the proposed algorithm works quickly and efficiently in sample scene.(2)Solve the problem of multi behaviors with single robot in conflicting behaviors. Among all the behaviors the robot may operate, some of which may be conflicting to each other. If one behavior is acted, the others cannot be executive at the same time by the same robot. Aiming at coordinating conflicting behaviors, we propose an algorithm for the coordination of multiple and possible conflicting behaviors selecting algorithm, devoted to reach the goal while collecting collectable objectives. The proposed endocrine based behavior selecting algorithm is composed of an artificial endocrine system, including two hormones, and three behaviors. The experiments are performed in simulation conditions. The results show that the algorithm proposed is able to coordinate a multiple and conflicting behaviors task, with less time and being more efficient.(3)Solve the problem of behaviors with multi robots, in task allocation. To work with the relation between energy consumption and working rate, we propose an artificial endocrine system to control the whole robot system. The system is composed of a pair of trophic/ inhibitory hormone, a kind of hormone and 3 correspondingbehavior controllers. The experiment result shows that the system will extend the working period of robot system and keep the system executing efficiently.At the end of the paper, some conclusions of the research work are made and some outlook of the next research direction or problems are discussed briefly.