individual
What campus are you from?
Daytona Beach
Authors' Class Standing
James Hand, Graduate Student
Lead Presenter's Name
James Hand
Faculty Mentor Name
Bryan Watson
Abstract
The translation between research and the real world can often be a daunting task, one typically left to dwell in the ‘future work’ section of a paper. This work presents the beginning of a transition from a custom-built AnyLogic multi-agent system simulation model to a physical robot platform hosted by the Georgia Institute of Technology. The Georgia Tech Robotarium is a globally accessible research support platform that enables the testing of multi-agent system control schemes. In addition, the Robotarium has a software simulator mimicking the hardware of the system in MatLab. Using this, the researcher tested insect inspired strategies for resisting infectious faults which had previously only been tested in a custom-built AnyLogic software simulator. This new testing approach showed key difference between the previous AnyLogic simulation and the Robotarium’s software simulator. Preliminary findings show distinct and unique differences between the two testing environments, with the best combination of strategies for resilience being different in each simulator. Initial analysis suggests these differences is due to the vastly different capabilities of agents within the simulations and the simulated environment’s size. What does hold true however, is a strong improvement of system performance and resilience with the use of Ant Inspired strategies when compared to the control in each simulation, up to 36% increase in swarm fault resistance. The next step will involve testing these same strategies on the physical Robotarium swarm.
Did this research project receive funding support from the Office of Undergraduate Research.
No
Testing Ant Inspired Faulted Agent Resilience Strategies using the Georgia Tech Robotarium
The translation between research and the real world can often be a daunting task, one typically left to dwell in the ‘future work’ section of a paper. This work presents the beginning of a transition from a custom-built AnyLogic multi-agent system simulation model to a physical robot platform hosted by the Georgia Institute of Technology. The Georgia Tech Robotarium is a globally accessible research support platform that enables the testing of multi-agent system control schemes. In addition, the Robotarium has a software simulator mimicking the hardware of the system in MatLab. Using this, the researcher tested insect inspired strategies for resisting infectious faults which had previously only been tested in a custom-built AnyLogic software simulator. This new testing approach showed key difference between the previous AnyLogic simulation and the Robotarium’s software simulator. Preliminary findings show distinct and unique differences between the two testing environments, with the best combination of strategies for resilience being different in each simulator. Initial analysis suggests these differences is due to the vastly different capabilities of agents within the simulations and the simulated environment’s size. What does hold true however, is a strong improvement of system performance and resilience with the use of Ant Inspired strategies when compared to the control in each simulation, up to 36% increase in swarm fault resistance. The next step will involve testing these same strategies on the physical Robotarium swarm.