In this work, we present a complete and robust aerial robot system for automated infrastructure inspection with a human machine interface, which is also satisfied with SWaP constrain. The complete and robust system consists of a state of art state estimation algorithm VINS-Fusion, a global consistency mapping method Surfel Fusion for spatial and temporal global optimized planning, a low-cost local occupancy grid map together with ESDF map for local replanning which can avoid obstacles in realtime. We highlight our work as a user-friendly system that can be used by everyone. Thus, besides suitable algorithms, the system is well designed in many aspects. Modularity hardware makes it hard to damage and easy to repair. Intelligence software and human machine interface allow the sys...[ Read more ]

In this work, we present a complete and robust aerial robot system for automated infrastructure inspection with a human machine interface, which is also satisfied with SWaP constrain. The complete and robust system consists of a state of art state estimation algorithm VINS-Fusion, a global consistency mapping method Surfel Fusion for spatial and temporal global optimized planning, a low-cost local occupancy grid map together with ESDF map for local replanning which can avoid obstacles in realtime. We highlight our work as a user-friendly system that can be used by everyone. Thus, besides suitable algorithms, the system is well designed in many aspects. Modularity hardware makes it hard to damage and easy to repair. Intelligence software and human machine interface allow the system could be used without any professional skills. Reasonable logic design, including autonomous airborne mapping with obstacle avoidance, designing mission path in an immersive simulator and autonomous inspection following an optimized trajectory, permits it to be utilized in inspection, search and rescue applications. Last but most important, a three-level safety design decreases the accident rate of the system. For testing our system, we first test the system module by module to verify the performance of each part. Then, test it integrally in the indoor and outdoor environment. It is also deployed to a real application and tested by several green hands, which makes the performance of our system more convincing.