Hybrid reasoning for teams of heterogeneous robots: finding an optimal feasible global plan

Official URL: https://www.aaai.org/ocs/index.php/SSS/SSS14/paper/view/7766

We address two key challenges of domains including teams of heterogeneous robots, hybrid reasoning for each team and finding an optimal global plan for multiple teams, by utilizing state-of-the-art automated reasoners. For hybrid reasoning, we propose (i) representing each team’s workspace taking into account capabilities of heterogeneous robots, (ii) combining different methods of integration to embed external computations (e.g., collision checks) into high-level representation and reasoning, (iii) further optimizing local feasible shortest plans by minimizing the total cost of robotic actions, where costs of actions can be defined in various ways. To find a shortest global plan, we propose a semi-distributed approach based on our earlier studies: we formulate the problem of finding an optimal coordination of teams that can help each other, prove its intractability, and describe how to solve this problem using automated reasoners. As a case study, we have formally represented a cognitive toy factory and showed an application of our hybrid reasoning and coordination approach on this domain. We have performed dynamic simulations and physical implementation of the cognitive toy factory utilizing KuKa youBots and Lego NXT robots.