Maja J Mataric´

Behavoir-Based Multi-Robot Coordination

We are pursuing a research program focused on studying, synthesizing, and analyzing robot group behavior and learning. The goal of the research is to understand the types of simple local interactions which produce complex and purposive group behaviors, both in nature and in synthetic systems. We have developed a bottom-up, behavior-based approach that can be used to structure and simplify the process of both designing and analyzing emergent group behaviors. The approach utilizes a biologically-inspired notion of basis behaviors for control, both at the individual and collective level.

Based on the notion of using a small set of composable behaviors as a substrate for control and representation, we have also developed methods for single and multi-robot learning. Our current research focuses on methods for developing adaptive multi-robot controllers for both loosely-coupled and tightly-coupled task domains. We are currently pursuing the following projects:

controller architectures and performance:

  • generating pareto-optimal and satisficing group behavior task domains:target tracking and hunting
  • port-based arbitration for distributed coordination; task domains:tracking multiple moving objects, formations
  • group-level dynamic load balancing; task domains:object transport and delivery in structured environments
  • adapting publish/subscribe communication paradigm for distributed task allocation; task domains:box pushing


  • learning models of interaction dynamics at different time-scales; task domains:distributed foraging and collection
  • dynamic run-time re-tasking using high-level behavior representations; task domain:object delivery
  • automatic detection of regime switching; task domains:foraging, collection, coverage

    biologically inspired models:

  • ant-inspired navigation strategies; task domains:object collection and large-scale navigation
  • ethologically-inspired variations on group organization; task domain:foraging

    interaction with non-robotic systems:

  • embedding robots on the Internet; task domains:exploration, mapping, delivery
  • human-robot interaction for task specification and learning; task domain:object delivery

    For details, see our projects page and the Interaction Lab home page.

    For relevant publications, please see Maja's publications page and the Interaction Lab publications page.

    This is a subset of our robot family used for group behavior and learning experiments. For a more complete view, see our facilities list.
    Go back to Maja's home page.

    Go to the Interaction Lab home page.