Autonomous Swarming Hierarchies (ASH)

Boston Fusion Corp. and the Cyber-Physical Systems Laboratory at Rutgers University propose Autonomous Swarming Hierarchies (ASH), a platform-agnostic multi-robot system (MRS) design software suite with three components: 1) a coordination module (CASH) that uses artificial intelligence/machine learning to automatically generate control policies for the robots comprising the system, 2) a networking module (NASH) that automatically synthesizes the MRS cyber topology, while enforcing connectivity constraints, modalities, encoding options, etc., and 3) an optimization module (OASH) that uses an optimization engine to search for the best physical and logical robot configurations that meet specific performance bounds. With CASH, a user can train an MRS to best coordinate the accomplishment of a mission; however, the user needs to specify a priori the types of robots in the system, their software configuration, and the system’s cyber topology. With CASH and NASH together, the user still needs to specify the types of robots that comprise the system, but the cyber topology is automatically synthesized. If OASH is also enabled, the types of robots or their software configuration will be optimized for the desired mission. ASH’s modularity affords maximum flexibility in MRS mission design and deployment.