Carnegie Mellon Robotics Institute
Robots are increasingly showing potential for support of human operations such as surveillance, reconnaissance and rescue. Robots can provide remote, distributed, real-time sensing and actuation while reducing the risks inherent to the human operators. In some domains, a team of robots can be more effective than the individual. A team can perceive its environment from multiple disparate viewpoints simultaneously. Moreover, it can distribute its resources among several members reducing the size, complexity and cost of the individual.
My work involves the control and coordination of a team of millimeter-sized, heterogeneous, resource-limited robots, called Millibots, which are designed to operate in unknown or partially known environments. Their small size gives them access to otherwise inaccessible areas while making them easier to conceal, deploy and manage. Because of the imposed size restrictions, individual Millibots are limited in the amount and types of sensing and processing they can carry. For this reason, we have chosen to develop the team in a modular, heterogeneous fashion. Based on the desired task, a team is configured with various sensing platforms. Current available platforms include; short and medium range sonar arrays, small video cameras, infrared ranging sensors and pyro-electric sensors.
Although robot teams may be designed to work in close conjunction with an operator, a degree of autonomy is necessary to free the operator from direct supervision and coordination. Robot teams must be able to operate in an unstructured environment with little or no previous information. One of the most significant skills a small team must master is the ability to autonomously explore its space. Given the limitations in power, processing and sensing, it must do so in an intelligent fashion. Because of the dependant nature of localization, mapping and navigation, I am developing algorithms and skills that allow the team to more effectively explore unknown spaces while maintaining global position in the context of the size and power budget. These algorithms include methods for fusing the different sensing modalities to produce maps and generating intelligent plans for exploration of unknown spaces.
|Research Interest Keywords|
|mobile robots, motion planning, multi-agent systems|
|The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.|
Contact Us | Update Instructions