National Science Foundation, NASA
Dubowsky, S., Argaez, D., Sunada, C., Moore, C., Madhani, A.,
Rutman, N., Cole, J.
The methods that were developed by this project include:
Force Workspace [41-45]:
This approach that is used in the design and motion planning of mobile robots maps constraints in the manipulator workspace such as joint limitations, force specifications, obstacles etc into the robot configuration space and forms contraint obstacles. Motion and system designs can be found using a search algorithm that do not violate these constraint obstacles.
Power Map [46-47]:
An analytical and graphical design and planning technique that provides the information required for analyzing and evaluating the power consumption of a class of multi-limbed mobile robotic systems.
Coordinated Jacobian Transpose Control [48-50]:
A method to control multiple control variables for both position and force of multi-limbed robotic systems in a unified and coordinated manner. This method can be interfaced very easily with higher level planners and controllers in a straightforward way using the concept of generalized variables.
Design of Modular Field Robots [51]:
The objective of this research has been the development of a framework for the automatic design of modular field robotic systems. It is based on the generation of computer-based hierarchical search over the possible robot assemblies and physics based rules to limit the search to "reasonable" assemblies.
Path-Planning of Field Robots Using Genetic Algorithms [52]:
A method to generate rapid plans for modular robotic systems used in unstructured field environments using the automatic programming technique of genetic programming.
Copyright 1995 by MIT.
All rights reserved.