Algorithms for Distributed Control

  • Distributed Computing and Optimizations. We have developed a Distributed Algorithm for Solving Linear Equations [pdf], which converges exponentially fast, operates asynchronously, work for time-varying networks and involve no step-size, followed by further improvement including elimination of initialization [pdf], and decrease of state dimension by the sparsity [pdf]. We have also recently developed a discrete-time distributed algorithm for least-square solution [pdf], and solutions with minimum L1 norm [pdf].
  • Resilience for Distributed Control. We aim to develop systematic approaches based on nodes' local information to achieve resilience against sophisticated cyber attacks without identification/isolation.
  • Integration of Machine Learning into Distributed Control.  We are also interested in applying reinforcement learning methods into distributed control of multi-agent networks. 

Experimental Work for Multi-Vehicle Coordination

Our lab is called 
AutoMous Lab . By deleting "no" from autonomous, we want to emphasize our "yes" capability to research related to autonomous.  "AutoMous" is composed of two parts, namely, Auto, which is our research object; and Mous, which suggests our goal is to build vehicles intelligent as mouse without eating! The research of AutoMous Lab  focuses on investigation of multi-vehicle coordination, for which we have achieved progress in the following topics.


 UAV Design for Energy Efficiency 
( BiCopter: VTOL + Horizontal Flying)

Abstract: The UAV built in our lab aims to achieve energy efficiency while enabling vertically take off and landing. Control of the UAV includes two phases, that is, propellers enables the UAV for VTOL (Phase I), and Propellers tilt to reach enough horizontal speed for fixed wings to generate aerodynamic lift (Phase II). 
 
 Collaborations of Multiple UAVs
(Autonomous flying/landing; vision-based control)

Abstract: A big UAV carries a small one for long distance travel. When they are close to target area, the small UAV autonomously takes off for missions, then autonomously lands back to the base UAV by only using its onboard cameras. 
 
  Distributed Control of Mobile Robots

Abstract:  We have developed a platform consisting of three iRobots for implementation of algorithms for formation control. Further applications into surveillance, coverage and exploration of the unknown is in progress. 
 
 
  Collaboration between UAV and Ground Vehicles
(In Progress)

Abstract: We are also interested in collaboration between smaller UAVs and ground vehicles. By introducing a UAV for a ground vehicle, the UAV could serve as an additional eye for the ground vehicle.