Multi-Vehcile Coordination and Control Under Network and Geographical Constraints
In this area of research I address multi-vehicle coordination and control under network management constraints. Under my supervision, we have created a control architecture and quering system that enables the realtime control and simulation of large groups of vehicles maintaining a network topology while navigating through vast topological areas with geographical constraints. These architecture combines reactive prioritized multi-task approaches with task-based planners. In the image below it is shown and Illustration of a UGV/UAV exploration scenario.



Funding source: The Boeing Company
Years: 2008-2009
Resources: $350,000
Research coordinator: Dr. Luis Sentis
Student: Mike Mintz (MS, CS)





























Click image to play video (Format: WMV, Size: 68MB)Click image to play video (Format: WMV, Size: 81MB)
Exploration behavior with disc formation: Number of vehicles is 100; grid squares are 1km in size; maximum network external disc coverage is 2km in diameter and minimum network separation is 1km; exploration pattern consists of concentric circles around the terrain; formation topology is a pack of minimal volume; brown obstacles are mountains, blue obstacles are lakes, and grey obstacles are cities; vehicle and group maximum speeds ares 50km/h; simulation speed is 360x; total grid area is 100km x 100km.Exploration behavior with ring formation: Number of vehicles is 50; formation topology is based on a ring potential attractor allowing higher swapping area per periodic circle; all other parameters are identical to the exploration behavior on the left.


OBOTICS
PUBLICATIONS
Luis Sentis, Mike Mintz, Arun Ayyagari, Craig Battles,  Susan Ying, Oussama Khatib, Control Methods for Multiple Vehicle Coordination with Network Management Constraints, The International Symposium of Robotics Research (ISRR), (in preparation, due April 2009).