Vanderbilt University

DARPA Mobile Autonomous Robot Software (MARS) Project Home Page


This is the home page of the Vanderbilt University Intelligent Robotics Lab Project funded under the DARPA Mobile Autonomous Robot Software (MARS) program.  Funding began July 1, 1999 and ended June 30, 2002.

The Vanderbilt team developed a software architecture that enables a mobile robot to learn by example and from experience, then share that knowledge with other robots.  This architecture is the computational foundation for a team of mobile robots that can take mission-level instructions from a human commander and proceed autonomously. Rather than a more traditional approach, we proposed to use an innovative agent-based software architecture, the Intelligent Machine Architecture (IMA), developed by the Center for Intelligent Systems (CIS) at Vanderbilt University. IMA affords operational flexibility and platform portability. It is used by several platforms in the CIS: a humanoid robot, ISAC (Intelligent Soft Arm Control), and several mobile robots, including ATRV-Jr (Scooter), Pioneer2-AT (Skeeter), HelpMate and several Trilobots.  Innovative concepts included the Sensory EgoSphere (SES), Spreading Activation Network-Reinforcement Learning (SAN-RL), Landmark EgoSphere (LES) among others.

We invite you to visit our DARPA-MARS Program Website and Learn more about Vanderbilt's mobile robot and humanoid robot research.

Kaz Kawamura, Principal Investigator