Inspection of critical infrastructure with drones is experiencing an increasing uptake in the industry driven by a demand for reduced cost, time, and risk for inspectors. Early deployments of drone inspection services involve manual drone operations with a pilot and do not obtain the technological benefits concerning autonomy, coordination, and cooperation. In this paper, we study the design needed to handle the complexity of an Unmanned Aerial System (UAS) to support autonomous inspection of safety-critical infrastructure. We apply a constructive research approach to link innovation needs with concepts, designs, and validations that include simulation and demonstration of key design parts. Our design approach addresses the complexity of the UAS and provides a selection of technology components for drone and ground control hardware and software including algorithms for autonomous operation and interaction with cloud services. The paper presents a drone perception system with accelerated onboard computing, communication technologies of the UAS, as well as algorithms for swarm membership, formation flying, object detection, and fault detection with artificial intelligence. We find that the design of a cooperative drone swarm and its integration into a custom-built UAS for infrastructure inspection is highly feasible given the current state of the art in electronic components, software, and communication technology.

Autonomous multi-drone navigation in unknown environments requires robust solutions for obstacle avoidance, formation cohesion, and inter-drone communication maintenance. We address these challenges using Riemannian Motion Policies (RMPs), a second-order dynamic framework for motion generation, within the RMPflow architecture. We design handcrafted and learned obstacle avoidance policies, including circle-based and LiDAR-based approaches, and introduce residual reinforcement learning as a novel mechanism for refining suboptimal handcrafted policies. To maintain reliable drone-to-drone communication, we propose a radio link quality policy that dynamically adapts the maximum permitted inter-drone distance based on signal strength, contrasting with a fixed worst-case baseline. These policies are composed within RMPflow into an efficient reactive motion planner and evaluated through 2D simulations across multiple configurations. Results show that residual learning substantially improves a suboptimal handcrafted policy, reducing collisions by 35.4% in drone formations and up to 74.7% for single drones. Notably, a policy trained without any initial obstacle avoidance capability failed to obtain such improvement, underscoring the importance of informed policy initialisation. The adaptive radio link policy achieved 20% to 25% shorter navigation time through randomly generated environments compared to a fixed maximum distance constraint, demonstrating the practical value of radio signal-aware formation control.