An architecture is presented for controlling robots that have multiple tasks, operate in dynamic domains, and require a fair degree of autonomy. The architecture is built on several layers of functionality, including a distributed communication layer, a behavior layer for querying sensors, expanding goals, and executing commands, and a task level for managing the temporal aspects of planning and achieving goals, coordinating tasks, allocating resources, monitoring, and recovering from errors. Application to a legged planetary rover and an indoor mobile manipulator is described.
A task control architecture for autonomous robots
1990-03-01
Conference paper
No indication
English
Task-level control for autonomous robots
NTRS | 1994
|Task-level control for autonomous robots
AIAA | 1994
|