Van Bien, Bui.. Teleoperation system using neural network based multiple model adaptive predictive control. Master's Degree(Mechanical Engineering (International Program)). King Mongkut's University of Technology North Bangkok. Central Library. : King Mongkut's University of Technology North Bangkok, 2007.
Teleoperation system using neural network based multiple model adaptive predictive control
Abstract:
Today, applications of master-slave teleoperation systems can find in many
areas from micro to macro scales. In a teleoperation system, the master is moved by a
human operator, and the slave is controlled to follow the motion of the master.
Environment model and communication time delays of teleoperation system are
usually variant, which will cause bad performance, even instability of the system. We
have developed a theoretical framework to design the neural network based multiple
model adaptive predictive controller to solve this problem. First, neural network
model based of any possible environment is built up. The teleoperation controller is
designed under the effects of time delay. The state observer of forward time delay is
built for all environment models to predict slave state. The model of total
teleoperation system is established. Finally, the control parameters can be given
conveniently using the result of transparency analysis. Therefore, the performance of
teleoperation system is worked properly.
The proposed control strategy is tested for verification system. In the
simulation, the stability and performance result of teleoperation system is achieved
under the proposed controller.