Reconfigurable Modular Manipulator System
Pradeep K. Khosla
Christiaan J.J. Paredis
The Robotics Institute
Carnegie Mellon University,
Pittburgh PA 15213.
Christiaan J.J. Paredis, and Pradeep K. Khosla, "RMMS: Reconfigurable Modular Manipulator System Project," in Video Proceedings of the 1997 IEEE International Conference on Robotics and Automation, Albuquerque, NM, April 19-24, 1997.
MPEG (size 7.24M)
The RMMS utilizes a stock of interchangeable joint (actuator) and link modules of different size and performance specifications. The modularity in mechanical, electrical and electronic design allows the user to design the optimal manipulator for the task at hand. The RMMS extends the concept of modularity to also include the control algorithms and task planning software. With this combination of capabilities, the RMMS can be configured to meet the task requirements as they arise at the application site. This modular design philosophy has several advantages over conventional robotic systems, such as economy of manufacture, ease of modification and repair. Some potential application areas of the RMMS are in construction, space, nuclear and manufacturing environments.
In order to fully exploit the potential of this new class of manipulators, the Advanced Manipulators Laboratory is addressing challenging theoretical and technological issues. We have already built a prototype RMMS that includes four joints and links of varying sizes. To control the RMMS, we have also built a real-time control system Chimera.
Current research is addressing the following issues:
|We have also developed modular mechanical, electrical power, and electronic data connectors. The main idea being that as soon as the modules are mated mechanically, the power and the data connections are also established. The modular data connectors will serve as the control bus during the operation of this system.|
Automatic Generation of Kinematics, Dynamics and Control. We have developed algorithms that read the description of the individual modules (from the embedded microprocessor) and automatically create kinematic, dynamic and control models. These algorithms serve to make the low level details transparent to the user.
Mapping Tasks into Manipulator Configurations. We are developing a methodology to map the kinematic and dynamic characteristics of a task (such as workspace, end-effector velocity and acceleration, accuracy) into the kinematic and dynamic configuration of the desired manipulator. This will allow us to determine the optimal manipulator configuration for the task at hand.
Architectures for Real-Time Control. We are also developing both software and hardware architectures for sensor-based control of manipulators. We have developed a real-time kernel Chimera for implementing hierarchical multiple sensor-based controllers.