RESEARCH ACTIVITIES

Primary interests are: embedded systems, wireless systems, autonomous robotics.

Secondary interests are: game theory, multi-agent systems

Current Activities:

Improving Embedded Systems Development, primarily with TI MSP430 and MicroChip PIC microcontrollers.

Also, working with Dr. Curtis Welborn of UVSC on a simplified force feedback system for use primarily on multi-legged robots that employ machine learning, using reinforcement learning techniques.

Previously, 2005, 2006: 

Application of Reinforcement Learning to Specific Classic Game Theory Problems

Investigated the application of computer based reinforcement learning (RL) methods to certain classic game theory problems. Economic game theory predicts the play or behavior of hedonistic agents. Computer based RL algorithms (SARSA) were applied to several classic economic game theory problems. The results from the RL agents was compared to the results predicted by economic game theory, in particular Nash Equilibrium conditions. The Centipede Game was the focus of this experimental study.  This work was presented at the IEEE Region 5 Technology and Science Conference in San Antonio, Texas in April 2006.

Achieving Cooperative Behavior Among Hedonistic Agents in Control Systems

Cooperative behavior can emerge from the actions of hedonistic multi-agents with very limited and simple one-way communications.    My research pursued minimizing inter-agent communications while achieving cooperative results.  The current approach is through computer simulation.  This work was presented at the IEEE Region 5 Technology and Science Conference in San Antonio, Texas in April 2006.