A time-optimal trajectory planning approach based on calculation cost consideration

Abstract

A new smooth online near time-optimal trajectory planning approach is proposed to reduce the computation cost compared to traditional dynamics trajectory planning method. The robot path is expressed by a scalar path coordinate s. The joints torque bounds and speed bounds are transformed into the s-s plane, which can generate the limitation curves of pseudo-velocity. The maximum pseudo-velocity curve which meets the limitation of torque and speed is constructed through the feature points and control points in the s-s plane using cubic polynomial fitting method. Control points used for cubic polynomial construction are optimized by the Golden-Section method. The calculation time of this method is shorter than traditional dynamics method. Compared to five-order polynomial trajectory optimization method, the approach proposed in this paper can effectively take advantage of joints maximum speed.