Communicating Mobile Robotics |
||
Small Robot Testbed |
||
|
 |
|
 |
||
Large Robot Testbed |
||
| One aspect of our research on
communicating mobile robots is aimed at understanding the use of
real-time distributed sensor data to achieve coordinated control of the
motions of groups of mobile robots. The problems addressed by this
research lie at the intersection of appliead nonlinear control theory
and real-time communications over bandwidth limited channels. A
principal aim of the research here is to develop a large catalogue of
simple controlled motions which in appropriate sequential combinations
permit autonomous nonholonomic vehicles to assemble themselves and
execute coordinated motions in highly structured formations. Click here for more information about the IML robots. Click here for leader-follower video(980K). |
 |
|
Free Space Optical Communication |
||
| In cooperation with Professor Bifano of the Manufacturing department of Boston University, we have invested effort in innovative ideas for mobile free-space laser Communication for autonomous vehicles. Free-space laser communication has been used to connect networks from tall buildings and spare the effort of laying cables. For the application with autonomous vehicles, free-space laser has advantages over radio frequency communication in that it can communicate over very long range with relatively low power; also that it is stealthy, which is important for certain tasks. A major challenge is how to let the transceivers point to each other before communication is established, since the mobile vehicles cannot tell each other whether they are receiving the laser beam in the process of establishing the link. The system we built use retro-refectors to provide feedback for the transceivers to establish the link. The figures show the real setup in our lab, and the system we envision. |
  |
|
Stochastic Agents in Robotic Surveillance |
||
| We consider a team of
autonomous mobile robotic agents engaged in a surveillance mission. It is desirable not the have the agents move in a predictable fashion so any invader or enemy can not plan their movements to avoid the surveillance agents. We are developing strategies in the general context of a developing theory of optimal control of mixing in finite state Markov Chains. The research emphasizes methods which minimize centralized computation and communication requirements by focusing on local rules for each agent. We investigate both the problem of surveillance coverage and intruder detection. |
 |
|
|
||
|
 |
Control of Mechanical Systems |
|
|
This research has a number of applications components
including motion planning and control of kinematically redundant
manipulators, problems associated with anholonomy in planning motions
for robots which have elastic joints and other components which store
energy, the control theory of fluid structure interactions, the
nonlinear control theory of microelectromechanism dynamics, adaptive
optics, and network mediated control of large scale device arrays.
|
 |