abstrac In this research a planning methodology for nonholonomic mobile manipulators in the presence of obstacles is developed. The method employs smooth and continuous functions such as polynomials. The proposed method includes extracting time history of motion of mobile robot. For sampling it is supposed obstacles can be enclosed in ellipse or circles. Platform that has been used in this research is differentially-driven platform that very usable. The core of the method is based on mapping the nonholonomic constraint to a space where it can be satisfied trivially. Combination of platform and manipulator causes robot operates in extended work space. However studying of such systems include kinematics redundancy that make more complicated problem, but gives more capabilities to robotic systems, because of the multiple solutions obtainable in specified workspace. One method of selection amongst the set of solutions is to choose a suitable criterion and to optimize it. In this research the problem of path planning with optimization of dynamic indices has been accomplished using genetic algorithm and direct search in order to find the global optimum. Obstacles existence adds kinematics constraint into optimization problem. In order to apply the kinematics constraint into the optimization problem, two methods have also been used. The first one consists of using the analytic solution of the kinematics equation to reduce dimension of the problem, transforming the constraint optimization process to an unconstraint one. The second one that has been used in this research, kinematics constraint has been added to dynamic indices using Lagrange multiplier and penalty parameter. Various simulation of platform with a 3-link manipulator are presented to show the effectiveness of the presented method.
