| Abstract: | This article studies the trajectory planning of redundant robots performing tasks within an enclosed workspace. Configuration control of kinematically redundant manipulators using the pseudo‐inverse with null‐space projection method is a well‐known scheme. One advantage of this method is that the gradient of an objective function can be included in the homogeneous term to optimize the objective function without affecting the position of the end‐effector. Using different objective functions, this method can achieve redundancy resolution such as obstacle or joint limits avoidance. Along this line of redundancy resolution, a switching objective function is proposed. We modify Liegeois' joint angle availability objective function so that the midpoints of each joint are switched at a series of prespecified key path points for the end‐effector to achieve. These key path points are planned beforehand according to the geometry of the constrained workspace. The trajectory planning problem can then be viewed as a series of proper postures (i.e., midpoints) determination problems at the key path points. The proper postures are determined using a combination of the potential field method and the elastic model method that takes into account joint operating ranges and the motion tendency of the end‐effector. A variable weighting technique to achieve the proper postures effectively is also presented. Simulations of a planar eight‐link robot in a constrained workspace illustrate the effectiveness of the switching objective function with the variable weighting approach in trajectory planning problems. ©1999 John Wiley & Sons, Inc. |
