Self-Healing Ground-and-Air Connectivity Chains Vivek Shankar Varadharajan, David St-Onge, Bram Adams and Giovanni Beltrame Abstract— The coordination of robot swarms – large de- centralized teams of robots – generally relies on robust and efficient inter-robot communication. Maintaining communica- tion between robots is particularly challenging in field de- ployments. Unstructured environments, limited computational resources, low bandwidth, and robot failures all contribute to the complexity of connectivity maintenance. In this paper we propose a novel lightweight algorithm to navigate a group of robots in complex environments while maintaining connectivity by building a chain of robots. The algorithm is robust to single robot failures and can heal broken communication links. The algorithm works in 3D environments: when a region is unreachable by wheeled robots, the chain is extended with flying robots. We test the performance of the algorithm using up to 100 robots in a physics-based simulator with three mazes and different robot failure scenarios. We then validate the algorithm with physical platforms: 7 wheeled robots and 6 flying ones, in homogeneous and heterogeneous scenarios. I. INTRODUCTION Swarm robotics is a field of engineering studying the use of large groups of simple robots (generally with low sensing and computation capabilities) to perform complex tasks [1]. Ideally, a single robot failure in a swarm does not compromise the overall mission because of the inherent redundancy of the swarm [2]. With robustness to individual failures and scalability, the deployment of robotic swarms is foreseen as cost effective solution for spatially distributed tasks over large areas such as exploration [3], search and rescue [4], and area coverage [5]. In many such applications, coordination among the robots is of vital importance and the ability to coordinate depends largely on the ability of the swarm to communicate. A reliable communication infrastructure allows the robots to exchange information at any time. However, in reality there are many sources of failures (environmental factors, wear and tear, etc) that can break connectivity and compromise the deployment. Consider the Fukushima accident in 2011: robots were deployed to inspect the collapsed nuclear power plant (transmitting a video feed), but were subject to extreme failure rates dues to radiation. In our previous work [6], we introduced a chain-forming algorithm in 2-dimensional uncluttered spaces. This paper extends the approach to complex tridimensional environ- ments and heterogeneous robot teams. We propose a de- centralized, failure-tolerant chain-forming algorithm that can Mr. Varadharajan, Dr. Adams and Dr. Beltrame are with the Com- puter and Software Engineering department, ´ Ecole Polytechnique de Montr´ eal, 2900 Boul ´ Edouard-Montpetit, Qu´ ebec, CA e-mail: (vivek- shankar.varadharajan@polymtl.ca). Dr. St-Onge is with the Department of Mechanical Engineering, ´ Ecole de technologie sup´ erieure, 1100 Notre-Dame St W, Montr´ eal, QC H3C 1K3 preserve a desired network topology while navigating a complex environment and completing mission goals. Our approach is lightweight, leaving processing power for other parallel tasks. In practice, we progressively use the available robots of a swarm to form a communication chain from a ground station to a target. We set the desired network structure by specifying the minimum number of links that should be maintained between the robots, target, and ground station. Our algorithm uses a path planner (e.g. RRT * [7]) to generate a viable path, and then uses this path as a template to build a chain of robots towards the target. The contributions of this work can be summarized as: 1) the mathematical formulation of the self- healing chain-forming algorithm, that progressively places the robots towards the target with a configurable network structure; 2) the performance evaluation with various level of injected failures in three simulated complex environments using a physics-based simulator; 3) the validation of the al- gorithm with a small swarm of 7 wheeled and 6 flying robots, in homogeneous and heterogeneous team deployments. This paper is organized as follows: we start with a brief summary of the related work in Section II, then we detail our model in Section III, and our proposed algorithm in Section IV. Finally, section V provides the experimental results. II. RELATED WORK The problem of preserving connectivity is widely dis- cussed in the literature, including several recent contribu- tions [8], [9], [10], [11]. Many of these approaches compute an algebraic connectivity metric [12]: a computationally intensive approach and sensitive to noise for distributed implementations [13]. A handful of other approaches exert virtual forces on the swarm members to preserve a pre- existing topology [14], [15], [16]. One approach is to design reactive control laws using connectivity as an imposed constraint. For example, in [17] two control laws were developed for rendezvous and forma- tion tasks, maintaining an initially connected configuration and effectively preserving connectivity. A similar approach was implemented in [18], achieving rendezvous among a group of agents and preserving an initial connected condition using a potential based controller. The approaches that fall into this category rely on global coordination, reducing their overall scalability, i.e. making them less appropriate for larger groups of robots. Other works are more explicitly dealing with dynamic connectivity as they implement control laws that maximize an algebraic measure of connectivity (i.e. the second eigen- value of the Laplacian of the robots’ connectivity graph). arXiv:1909.10496v1 [cs.RO] 23 Sep 2019