TRACLabs engineers have developed intelligent control architectures that provide robust autonomy to robotic platforms. These architectures combine reactivity with deliberation to accommodate both real-time closed-loop control and long-range planning and execution of tasks. Our approach emphasizes adjustable autonomy whereby a human operator can easily intervene at any level of robot operation. TRACLabs personnel have extensive experience with control of highly dynamic manipulation systems.

Layered Control Architectures
TRACLabs Inc. researchers invented the layered approach to controlling robotic systems. This software is called 3T because it has three interacting layers (or tiers) of control – one for low-level robot control, one for sequencing basic robot operations and one for planning robotic resource and time constraints.

Sensor-To-Symbol Architectures
A sensor-to-symbol architecture regularizes the connections between the sensed physical world and the symbols that represent that world. TRACLabs Inc. is building, under DARPA contract, a sensor-to-symbol architecture. The architecture can watch various sensory streams and match the outputs of simple or complex sensory algorithms onto ontological classes.

Key Publications:

• R. P. Bonasso, R. J. Firby, E. Gat, David Kortenkamp, D. Miller, and M. Slack, Experiences with an Architecture for Intelligent, Reactive Agents, Journal of Experimental and Theoretical Artificial Intelligence , Vol. 9, No. 1, 1997.
• David Kortenkamp, Robert Burridge, R. Peter Bonasso, Debra Schrenkenghost, and Mary Beth Hudson, An Intelligent Software Architecture for Semi-autonomous Robot Control, 3rd International Conference on Autonomous Agents Workshop on Autonomy Control Software, May 1999.
• David Kortenkamp and Reid Simmons, Robotic Systems Architectures and Programming, Springer Handbook of Robotics, Springer-Verlag, 2008.