Model-Based Control of a Robot Manipulator

摘要: Model-Based Control of a Robot Manipulator presents the first integrated treatment many most important recent developments in using detailed dynamic models robots to improve their control. The authors' work on automatic identification kinematic and parameters, feedforward position control, stability force trajectory learning has significant implications for improving performance future robot systems. All main ideas discussed this book have been validated by experiments direct-drive arm.The addresses issues building accurate applying them high It describes how three sets - model links inertial rigid-body loads can be obtained automatically experimental data. These are then incorporated into single learning, MIT Serial Link Direct Drive Arm, which these were developed applied is one few manipulators currently suitable testing such concepts.Contents: Introduction. Arms. Kinematic Calibration. Estimation Load Inertial Parameters. Feedforward Computed Torque Control. Learning. Dynamic Stability Issues Force Conclusion.Chae An Research Staff Member, IBM T.J. Watson Center, Christopher Atkeson an Assistant Professor John Hollerbach Associate Department Brain Cognitive Sciences Artificial Intelligence Laboratory. included Series edited Patrick Winston Michael Brady.