Innate Motivation for Robot Swarms by Minimizing Surprise: From Simple Simulations to Real-World Experiments

Applications of large-scale mobile multirobot systems can be beneficial over monolithic robots because of higher potential for robustness and scalability. Developing controllers for multirobot systems is challenging because the multitude of interactions is hard to anticipate and difficult to model. Automatic design using machine learning or evolutionary robotics seem to be options to avoid that challenge, but bring the challenge of designing reward or fitness functions. Generic reward and fitness functions seem unlikely to exist and task-specific rewards often have undesired side effects. Approaches of so-called innate motivation try to avoid the specific formulation of rewards and work instead with different drivers, such as curiosity. Our approach to innate motivation is to minimize surprise, which we implement by maximizing the accuracy of the swarm robot's sensor predictions using neuroevolution. A unique advantage of the swarm robot case is that swarm members populate the robot's environment and can trigger more active behaviors in a self-referential loop. In this article, we summarize our previous simulation-based results concerning behavioral diversity, robustness, scalability, and engineered self-organization, and put them into context. In several new studies, we analyze the influence of the optimizer's hyperparameters, the scalability of evolved behaviors, and the impact of realistic robot simulations. Finally, we present results using real robots that show how the reality gap can be bridged.

Automated summary

Applications of large-scale mobile multirobot systems have advantages in terms of robustness and scalability. Developing controllers for these systems is challenging, and automatic design using machine learning or evolutionary robotics can be a solution. However, the challenge lies in designing reward or fitness functions. Innate motivation approaches, such as minimizing surprise, can be used to avoid specific rewards and work with different drivers like curiosity. The unique advantage of swarm robot cases is that swarm members can trigger more active behaviors in a self-referential loop.