An Optimal Control Approach to Mapping GPS-Denied Environments Using a Stochastic Robotic Swarm

摘要: This paper presents an approach to mapping a region of interest using observations from robotic swarm without localization. The robots have local sensing capabilities and no communication, they exhibit stochasticity in their motion. We model the population dynamics with set advection-diffusion-reaction partial differential equations (PDEs). map environment is incorporated into this spatially-dependent indicator function that marks presence or absence throughout domain. To estimate function, we define it as solution optimization problem which minimize objective functional based on temporal robot data. performed numerically offline standard gradient descent algorithm. Simulations show our can produce fairly accurate estimates positions geometries different types regions unknown environment.