摘要：Existing rigid robot arms are often designed to have two or three joints in view of costs and manufacturing points. However at an initial design stage, the number of joints and lengths of arms should be designed more freely and adequately from the points of total efficiency and costs. This study aims at developing and providing a general and convenient method to obtain optimum number of joints and lengths of a rigid robot arm at initial design stage. The design of optimum joints and lengths, of course, depends upon trajectories, obstacles in working space, tasks and other many factors. At an initial design stage it is not so easy to determine optimum number of joints and lengths under consideration such many factors but is worthwhile to study on seeking for a possible method. Hence this study proposes a general design method where a trajectory itself is taken as design variables beside number of joints and lengths of robot arms because a trajectory is considered as one of prominent factors to effect on. such two measures energy and manipulability are adopted as objective functions. Then an optimum problem is formulated as multi-objective optimum design problem. This problem is difficult to solve by conventional optimization methods and Genetic Algorism (GA) is made use of. Based to the proposed method, numerical examples in 2- and 3- dimensional spaces are demonstrated. Model experiments are also carried out. All those results are examined and show the validity and effectiveness.