Dynamic Trajectory Evolution Incorporating Modeling Of Arm Movement Neural Data
S. Konyk · 2005
An approach is presented for the real-time generation and tracking of manipu- lator jdnt trajectories. The trajectory gener- ation is derived from a vector-integration-to- endpoint (VITE) model of human arm neural dynamics. The VITE model ia coupled with models of manipulator kinematics and dynam- ics. The joint trGectories are implicitly gen- erated and converge to the desired end effec- tor global coordinates without requiring an ex- plicit solution to the inverse mapping problem. The manipulator kinematics model assures re- alizable trajectories, while the manipulator dy- namics model regulates the speed of trajectory evolution to attain satisfactory tracking accu- racy. The resulting manipulator motion implice My exhibits VITE model characteristics inclnd- ing synchronous jdnt operation, no overshoot and bell shaped velocity profiles. The method is illustrated with a two-link manipulator.