Abstract:Swarm robotic systems are systems in which multiple robots having simple functionality perform tasks through their cooperation, and are advantageous in that they can exhibit non-trivial macroscopic functions such as adaptability, fault tolerance, and scalability. We previously proposed a simple model of swarm formation inspired by friendship formation process in human society, and demonstrated via simulation that various non-trivial patterns emerge. In this study, we examine the applicability of the proposed model to a swarm robotic system. As a first step, we developed five robots and demonstrated via real-world experiments that the simulation results can be largely reproduced.
Abstract:In this paper, we introduce Path Integral Networks (PI-Net), a recurrent network representation of the Path Integral optimal control algorithm. The network includes both system dynamics and cost models, used for optimal control based planning. PI-Net is fully differentiable, learning both dynamics and cost models end-to-end by back-propagation and stochastic gradient descent. Because of this, PI-Net can learn to plan. PI-Net has several advantages: it can generalize to unseen states thanks to planning, it can be applied to continuous control tasks, and it allows for a wide variety learning schemes, including imitation and reinforcement learning. Preliminary experiment results show that PI-Net, trained by imitation learning, can mimic control demonstrations for two simulated problems; a linear system and a pendulum swing-up problem. We also show that PI-Net is able to learn dynamics and cost models latent in the demonstrations.