Proceedings in Manufacturing Systems, Volume 8, Issue 2, 2013 ISSN 2067-9238 AN ANALYSIS OF THE ROBOT COLLISION AVOIDANCE USING THE PROGRAMMING THROUGH IMITATION Aurel FRATU 1,* , Michel DAMBRINE 2, 3 1) Prof., PhD, Dept. of Automatics, Electronics and Computers, „Transilvania” University of Braşov, Brasov, Romania 2) Prof., PhD, Univ Lille Nord de France, F-59000 Lille, France, 3) UVHC, LAMIH, F-59313 Valenciennes, France Abstract: This paper presents an analysis of the collision avoidance of the cooperative robots using the programming through imitation. Each physical robot acts fully independently, communicating with cor- responding virtual prototype and imitating her behavior. Each physical robot reproduces the motion of her virtual prototype. The estimation of the collision-free actions of the virtual cooperative robots and the transfer of the virtual joint trajectories to the physical robots who imitate there virtual prototypes, are the original ideas. We tested the present strategy on several simulation scenarios, involving two virtual ro- bots and estimating collision-free actions, during of the cooperative tasks. Key words: virtual robots, cooperative robots, collision avoidance, motion imitation. 1. INTRODUCTION 1 Important advancements were produced in the last years in fields of intelligent robots. From this point of view, promising applications were developed in robot- robot cooperation. Cooperative robots are permanently in danger to be in collision. Therefore installations with cooperative robots in real world, require collision avoidance methods, which take into account the mutual constraints of the robots. A key requirement for cooperative efficient operation is good coordination and reciprocal collision avoidance. The contact of the robot with an obstacle must be de- tected and it will cause the robot to stop quickly and thereafter back off to reduce forces between the robot and environment. The problem of the contact with obsta- cle imposes the null velocity in the moment of the impact and to obtain the zero-velocity points on the pathway. The collision detection simply determines if two geomet- ric objects are intersecting or not. The intersecting of two objects is possible in the virtual world, where the virtual objects can be intersected and there no exist the risk to be destroyed. Using this strategy one detects collisions in all direc- tions, protecting not only the physical end-effectors but also the work pieces and the physical robot itself. The ability of predicting of the behavior of coopera- tive robots is important in design; the designers want to know whether the robot will be able to perform a typical task in a given time frame into a space with constraints. The control engineer cannot risk a valuable piece of equipment by exposing it to untested control strategies. * Corresponding author: Str. Mihai Viteazu nr. 5, Corp V, et III Cod. 500174, Jud. Braşov, Romania Tel.: +40 268 418 836 E-mail addresses: fratu@unitbv.ro (A. Fratu), michel.dambrine@univ-valenciennes.fr (M. Dambrine) Therefore, a facile strategy for contact detection and col- lision avoidance, capable of predicting the behavior of a robotic manipulators, becomes imperative. When the robots need to interact with their surround- ing, it is important that the computer can simulate the interactions of the cooperative participants, with the pas- sive or active changing environment in the graphics field, using virtual prototyping. In this paper, we propose a fast method that simulta- neously determines actions for two virtual robots that each must cooperate with other. The actions for the cooperative tasks are computed for each virtual robot and are transferred, with a central coordination to corresponding physical robot which must imitate her virtual homonym. Thus, we will prove that our method guarantees the collision-free motion for each of the cooperative robots. In this paper we develop a formally analyze of a new collision avoidance strategy for a group of two coopera- tive robots. We assume that our strategy is able to deduce the exact shape, position and velocity of the virtual ob- stacles and of the virtual robots, in the virtual environ- ment. We transfer the behavior of the virtual cooperative robots, in the real world, to the physical cooperative ro- bots. This paper is focused on the collision avoidance through transfer the motion mapping from virtual space, in 3-D dimensional real space. 2. OVERVIEW OF IMITATION Imitation is an important learning mechanism in many intelligent systems including robots. It is easy to recuperate kinematic information from virtual robot mo- tion, using for example motion capture. Imitating the motion with stable robot dynamics is a challenging re- search problem [7].