Force control of heavy lift manipulators for high precision insertion tasks

Abstract

The inherent strength of robotic manipulators can be used to assist humans in performing heavy lifting tasks. These robots reduce manpower, reduce fatigue, and increase productivity. This thesis deals with the development of a control system for a robot being built for this purpose. The task for this robot is to lift heavy payloads while performing complex insertion tasks. This task must be completed on the deck of a naval vessel where possible disturbances include wind, rain, poor visibility, and dynamic loads induced by a swaying deck. The primary objective of the controller being designed here is to allow for insertion of the payload despite tight positioning tolerances and disturbances like surface friction, joint friction, and dynamic loads from ship motions. A control structure designed for intuitive interaction between the robot and operator is analyzed and shown to be stable using an established environment interaction model. The controller is shown to perform within established specifications via numerical simulation based on simple user inputs. An additional objective of this controller design is to prevent part jamming during the insertion task. With a large, powerful manipulator, the chances of a jam occurring is high. Without the use of bilateral force feedback, it will be difficult for the operator feel when these jams will occur and there will be no information about how to prevent them. This thesis analyzes the geometry and mechanics of the jamming problem and derives a control system to assist the user in preventing these jams. These methods can be extended to other insertion tasks simply by specifying the appropriate geometry.