A distributed simplex architecture for multi-agent systems

We present the Distributed Simplex Architecture (DSA), a new runtime assurance technique that provides safety guarantees for multi-agent systems (MASs). DSA is inspired by control architecture of Sha et al., but with some significant differences. The traditional approach limited to single-agent or MAS centralized scheme. addresses this limitation extending scope include MASs under distributed control. In DSA, each agent runs local instance such preservation in instances implies entire MAS. Control Barrier Functions (CBFs) play critical role. They are used define DSA’s core components—the baseline controller and decision module’s logic switching between advanced control—and they provide basis proof safety. general provided CBF-related optimization problem solved feasible (has solution) at time step which also propose novel extension designed avoid states infeasible. Finally, we experimental results several case studies, including flocking collision avoidance, safe navigation ground rovers through way-points, operation microgrid.