Charles River Applies Autonomous Learning to Swarm Tactics

Charles River Analytics Inc. has received funding from DARPA under the OFFensive Swarm-Enabled Tactics (OFFSET) program to develop Meta-Reinforcement Learning Innovation for Robust Swarm Tactics (MERLIN-RST). MERLIN applies a meta-reinforcement learning approach to discover and learn novel swarm tactics.

Through biology-inspired algorithms and deep machinelearning, tactics in the OFFSET program help swarms achieve and adapt to missionobjectives. However, human programmers cannot support the evolution of thesetactics quickly or accurately enough—numerous variants in adversary strategyand evolving environments make manual development a challenge that may lead tosystem failures or brittle tactics.

“MERLIN explores a rich, complex, and unpredictable searchspace that describes the many possible forms a tactic can take andautomatically optimize it to a wide range of mission goals,” said MichaelHarradon, Scientist at Charles River Analytics and Principal Investigator onthe MERLIN effort. “We’re laying the groundwork for a highly flexible designprocess that can dramatically improve swarm effectiveness.”

Under the MERLIN effort, the company is augmenting OFFSETsimulators to model swarm performance in a wide range of challenging urbanenvironments as part of the OFFSET Program’s Fourth Swarm Sprint. MERLIN buildson its recent successes under the OFFSET program. With SATURN, Charles Riverdeveloped capabilities that give heterogeneous swarms of unlimited sizeresilient behavior while achieving mission objectives. Its EUROPA effort helpsoperators better control swarms in urban operations by providing novel,multimodal user interfaces tailored to tactical operational environments.