A unified framework for external wrench estimation, interaction control and collision reflexes for flying robots

Flying in unknown environments can lead to unwanted collisions with the environment. If not being accounted for, these may cause serious damage to the robot and/or its environment. Fast and robust collision detection combined with safe reaction is therefore essential in this context. Deliberate physical interaction may also be required in some applications. The robot can then switch into an interaction mode when contact occurs. The control loop must also be designed with interaction in mind. To implement these mechanisms, knowledge of environmental interaction forces is required. In principle, they may be measured or estimated. In this paper, we present a novel model-based method for external wrench estimation in flying robots. The estimation is based on proprioceptive sensors and the robot's dynamics model only. Using this estimate, we also design admittance and impedance controllers for sensitive and robust physical interaction. We also investigate the performance of our collision detection and reaction schemes in order to guarantee collision safety. Upon collision, we determine the collision location and normal located on the robot's geometric model. The method relies on the complete wrench information provided by our scheme. This allows applications such as tactile environment mapping.