Autonomous multiagent systems is an area that is currently receiving increasing attention in the communities of robotics, control systems, and machine learning (ML) and artificial intelligence (AI). It is evident today; how autonomous robots and vehicles can help us shape our future. Teams of robots are being used to help identify and rescue survivors in case of a natural disaster for instance. There we understand that we are talking minutes and seconds that can decide whether you can save a person's life or not. This example portrays not only the value of safety but also the significance of time, in planning complex missions with autonomous agents.
In this talk, I will go over our recent work which aims to develop a generic, composable framework for a multiagent system (of robots or vehicles), which can safely carry out time-critical missions in a distributed and fully autonomous fashion. The goal is to provide formal guarantees on both safety and finite-time mission completion, thus, to answer the question: “how trustworthy is the autonomy of a multi-robot system in a complex mission?” We refer to this notion of autonomy in multiagent systems as assured or trusted autonomy, which is currently a very sought-after area of research, thanks to its enormous applications in autonomous driving for instance. An important aspect of our work is its special emphasis on fast i.e., computationally tractable solutions for this otherwise very challenging planning problem.
In the first part of the talk, using tools from control theory (optimal control), formal methods (temporal logic and hybrid automata), and optimization (mixed-integer programming), I will describe two variants of (almost) real-time planning algorithms, which provide formal guarantees on safety and finite-time mission completion for a multiagent system in a complex mission. Our proposed framework is hybrid, distributed, and inherently composable, as it uses a divide-and-conquer approach for planning a complex mission, by breaking it down into several sub-tasks. This approach enables us to implement the resulting algorithms on robots with limited computational power, while still achieving close to real-time performance. We have validated the efficacy of our method on several use cases; I will mention two such examples during my talk, that are: autonomous search and rescue with a team of UAVs, and planning UAV-based inspection tasks in an industrial environment.
In the second part, I will briefly describe how we can translate and adapt these algorithms to safely learn actions and policies for robots in dynamic environments, so that they can accomplish their mission even in the presence of an uncertainty. I will introduce the ideas of self-monitoring and self-correction for agents using hybrid automata theory. Self-monitoring and self-correction refer to the problems in autonomy where the autonomous agents monitor their performance, detect deviations from normal or expected behavior, and learn to adjust both the description of their mission/task and their performance online, to maintain the expected behavior and performance. In this setting, we propose a formal and composable notion of safety in learning for autonomous multiagent systems, which we refer to as safe learning.
To round up my talk, I will briefly go over some of my other projects in the general areas of robotics and control, that I have been involved in during my time with Industry, and also during my masters at KAUST, and undergraduate at PIEAS.