Landing gear plays an important role in enabling unmanned aerial vehicles to engage in complex missions and protect expensive airborne sensors. Most existing landing gears are designed to satisfy only specific maneuvers in predefined scenarios. Although effective in predefined scenarios, these landing gears cannot guarantee the adaptiveness in multiple scenarios. A promising solution is to build a model automatically generated based on the actual requirements, so that the expected model structure can be obtained through a constraint generation process. However, translating the actual requirements into reasonable design constraints is complex. For this purpose, we propose a landing structure design process in the bistable space of the Kresling tube. Furthermore, we use a combination of graph search and controller to optimize the landing structure. We validate the bistatic and graph search combined design structure in a realistic environment and test the mechanisms flexibility and assisted landing capability. Experiments show that the structure combining spatial bistability with a graph search approach can safely and successfully land a quadrotor under various conditions while enhancing the UAV's shock absorption during landing.
Read full abstract