Title: Optimization of Polynomial Trajectories for Robotic Manipulators Abstract approved:_

In this thesis, a method is presented to construct minimum-time robot trajectories for predefined Cartesian end-effector path in a workspace containing obstacles. The method is preferably applied to a geometric collision-free path of a SCARA robot by using the theories of Bezier, B-spline, and parabolic blending curves. The motion of the manipulator is defined by Cartesian control points and represented by a sequence of generated Cartesian knots along the end-effector path which is transformed into sets of joint displacements with one set for each joint. Cubic spline functions are then used to fit the sequence of joint displacements for each joint. The minimum-time trajectory planning problem is formulated as the problem of minimizing the total traveling time subject to constraints on joint positions, velocities, accelerations, jerks, torques and end-effector acceleration. The modified pattern search goal programming method is proposed to solve this nonlinear optimization problem. The computer program, ROBOPATH, has been written to implement the algorithm for a manipulator with two links and two degrees of freedom. The examples show the algorithm to be a useful tool in the design of manipulators, robot tasks and workcells. The results show that different locations of a path and different path shapes resulted in different total traveling times. Also, as a result, approximation curve techniques gave shorter total traveling time than the interpolation curve technique