MARS
Short for Microprocessor Architecture Research Society, MARS at the Georgia Institute of Technology was established in 2002 by Hsien Hsin Lee. It is a part of DARPA (Defense Advanced Research Project Agency)-sponsored robotic research at the university, whose main concern is with robotic teams of Unmanned Air Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs).
MARS’ objective is to explore, research and bridge circuits, architecture, compiler, and software methods. It is associated with the Center for Experimental Research in Computer Systems (CERCS), and also affiliated with Computer Arch Research at Tech (Comparch).
MARS is also involved in various projects, such as the multi-level learning in hybrid deliberative or reactive mobile robot architectural software systems. And to carry out complex adaptive autonomous mobile robot systems in vibrant, hostile environments, MARS has developed an innovative learning technique in the framework of established DARPA-developed software.
MARS has also incorporated various learning components into their MissionLab software: the probabilistic situational recognition and indexing into behavior sets for opportunistic planning and reaction, exploiting case-based reasoning or CBR for situation-dependent behavior gain and assemblage switching at the reactive level of implementation, the utilization of CBR for reasoning in FSA (Finite State Automata) plan generation, specialized back up learning methods in order to reset behavior gaits at run-time, and the addition of Q-learning techniques for behavioral assemblage selection at a level over gain adjustment.
