High Performance Manipulator Control



Korean Electricity and Power Corporation (KEPCO)

Electricite de France (EDF)

Foster Miller Corp.

National Institutes of Health (NIH), via the MGH Northeast Proton Therapy Center

Principal Investigator

 Professor S. Dubowsky

Group Members

Dr. Steven Dubowsky

Marco Meggiolaro, PhD student

Problem Statement

Large robot manipulators are needed in field, service and medical applications to perform high accuracy tasks. Examples are manipulators that perform decontamination tasks in nuclear sites, space manipulators such as the Special Purpose Dexterous Manipulator (SPDM) and manipulators for medical treatment. In these applications, a large robotic system may need to have very fine precision. Its accuracy specifications may be very small fractions of its size.

Achieving such high accuracy is difficult because of the manipulatorís size and its need to carry relatively heavy payloads. Many industrial robots are unable to perform highly demanding maintenance tasks, due to fundamental physical limitations, such as joint friction. Further, many tasks, such as space applications, require systems to be lightweight so that structural deformation errors may become relatively large.

The goal of this research is to study novel methods for improving the performance of industrial robots under position and force control. Other system effects, such as base flexibility, are also investigated. (SB, SM, PhD)


1.    The Design and Control of High Performance Robotic Systems for the Nuclear Industry

Robotic systems could do important tasks in the hazardous or difficult to reach places common to the nuclear industry. However, the capabilities of currently available commercial robotic systems are limited compared to the task requirements of the nuclear industry. These tasks require a manipulator to exert large forces and to transport heavy loads with high accuracy position and force control. Control problems arise in a strong manipulator due to actuators with high friction and nonlinear dynamic behavior. It has been shown that control methods using joint torque sensors produce significantly improved performance. However, joint torque sensors are not practical for most systems. In this work, alternative sensory techniques to improve manipulator performance are being investigated.

More detailed description of the project.


2.    Error Analysis and Control for the Patient Positioning System

The Patient Positioning System (PPS) being designed for the new Northeast Proton Therapy Center at the Massachusetts General Hospital must reliably obtain extremely high accuracies for the positioning of a patient undergoing cancer therapy. In this project, error analysis has been conducted to determine the performance of the system. A method to identify the positioning end-effector errors of large manipulators has been developed. The method can identify the sources of the end-effector errors, both geometric and elastic errors. Previous calibration techniques didnít explicitly consider the wrench at the end-effector to compensate for elastic errors. The accuracy of the PPS was improved from the inherent 5-7mm to less than the required accuracy of 0.5 mm.

More detailed description of the project.

Copyright 1995-2001 by MIT.

All rights reserved.