Manipulation tasks with a dual arm system including obstacles removing Carlos Rodr´ ıguez Andr´ es Monta ˜ no Ra´ ul Su´ arez * Abstract The paper deals with the problem of planning move- ments of a two-hand system, considering the possibility of using one hand to remove potential obstacles in order to grasp a desired object with the other hand. The ap- proach is based on a Probabilistic Road Map that does not rule out samples implying collisions with removable objects but instead classify them according to the collided obstacle(s), and allows the search of free paths with the in- dication of which objects must be removed from the work- space to make the path be actually valid. The approach has been implemented and different tests were performed with considering a real two-hand robotic system with one hand in charge of grasping a desired object and the other in charge of removing the potential obstacles. Some run- ning examples both in simulation and a real workcell are presented in the paper using simulations and real experi- mentations. 1 Introduction Moving objects around is a problem of significant rel- evance in the application of robots, both in industrial and service robotics. It involves several other associated prob- lems, among which there are two main ones: the deter- mination of a proper grasp configuration for the available arm and gripper (considering aspects like the shape of the object and the task to be performed), and the determina- tion of collision free paths to arrive to a grasp configura- tion and to move the object from its initial configuration to a desired one; grasping and path planning are already classic problems in robotics. In this work we deal with the second problem under the following context: two robots are available (see Fig. 1), we want to grasp a particular object with one of the robots, there may be other removable objects in the environment acting as obstacles, and we can use the second robot to remove the obstacles if it is necessary. Note that this is a * The authors are with the Institute of Industrial and Control Engineering (IOC) - Polytechnic University of Catalonia (UPC), Barcelona, Spain (carlos.rodriguez.pacheco, andres.felipe.montano, raul.suarez@upc.edu). This work was partially supported by the Spanish Government through the projects DPI2010-15446 and DPI2011-22471. Figure 1. Dual arm system. frequent problem in everyday life for the human beings, and so will be for humanoids robots. The proposed approach simultaneously solve the prob- lem of finding the robot movements to grasp and move the desired object, determining which are the objects acting as obstacles that must be removed, and the problem of find- ing the robot movements to remove them. The approach allows the consideration of different grasping configura- tions for each object and selects the ones that allow a real solution to the stated problems. The paper is organized as follows. After this intro- duction, Setion 2 presents a review of related works and Section 3 presents the proposed approach, giving first an overview and then a formal description including the proper algorithms. Then, Section 4 deals with the imple- mentation and presents some application examples in sim- ulations and real experimentations and, finally, Section 5 summarizes the work and presents some topics deserving future work. 2 Related Work During the last two decades there has been significant research work concerning robot motion planning prob- lems. Quite effective general strategies have been de- veloped using sampling-based techniques, and among the most relevant approaches are the Probabistic Road Map planners (PRM) [6] and Rapidly-exploring Random Trees planners (RRT) [9]. In order to speed up query path