1989 Fiscal Year Final Research Report Summary
A humanic controller for a maipulator arm via automatic skill acquisition from human experts
| Project/Area Number |
63460107
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
機械力学・制御工学
|
|---|
| Research Institution | KYOTO UNIVERSITY |
|---|
Principal Investigator |
TOKUMARU Hidekatsu Kyoto University, Faculty of Engineering, Professor, 工学部, 教授 (50025813)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
IIGUNI Youji Kyoto University, Faculty of Engineering, Assistant Professor, 工学部, 助手 (80168054)
ISHIDA Yoshiteru Kyoto University, Faculty of Engineering, Assistant Professor, 工学部, 助手 (80159748)
|
|---|
| Project Period (FY) |
1988 – 1989
|
| Keywords | Impedance control / Control rule / Hybrid control / Intelligent control / Simulation / Signal processing / Artificial intelligence / Divect dvive robot |
|---|
| Research Abstract |
An approach to the teaching of skills that human workers have acquired in performing specific tasks is presented, and a humanic controller for a manipulator arm is designed.
First, a skilled worker shows how to perform a given task, and his motions are measured. The force exerted by the worker and the displacement of the tool are analyzed and the relationship is derived from the data by using a curve fitting technique. The identified relationship is used as a reference model for controlling a manipulator arm to replicate the expert motion. The impedance with which a human expert holds the grinding tool is identified and transferred to the robot controller so that the system may follow the impedance model. Secondly, an efficient method is developed to generate programs for the hybrid position/force control of robots from teaching data.
An operator's motion is measured in terms of the force exerted by the operator and the position of the end-effector. The acquired data are then analyzed to
… More
understand what the operator intended to do, and the hybrid control program is generated. Determinations are made of which control mode, position or force, should be taken in each direction, how much force should be exerted, and what trajectory the end-effector should follow. The interpreted motion is then translated into a robot language, which explicitly describes the motion strategy that the human operator conceived. Lastly, the teaching of manipulative skills is developed. Teaching data acquired from a human expert are processed on a computer to attain;his skillful manipulation strategies.
The strategies are described by a group of control laxs that relate sensor signals to motion commands. Sensor signals are processed by using pattern recognition techniques to interpret sensor information and to allow the robot to recognize the state of the process.The control laws designate which control action the robot should take in response to each signal pattern generated in the process. It is demonstrated that the robot can mimic the skillful manipulation of the human expert and perform the task efficiently. Less
|
