Autonomous Navigation of Quadrotors Using Tactile Feedback

Abstract

In this paper, we present a novel approach for autonomous navigation of quadrotors in complex unknown environments using tactile feedback. The approach uses an array of force/contact sensors on the quadrotor to determine local obstacle geometry and follow contours of sensed objects. The approach is particularly useful in scenarios where visibility is limited, such as in dark or smoky/foggy conditions, in which vision-based navigation is not possible. To show the efficacy of the proposed approach, we perform simulation studies in a variety of environments and demonstrate that the quadrotor is able to autonomously navigate without any a priori knowledge of the environment and without relying upon any vision-aided sensing of the environment.