A Central Pattern Generator-Based Control Strategy of a Nature-Inspired Unmanned Aerial Vehicle

This work introduces a Central Pattern Generator (CPG)-based control formulation for a dual-winged nature-inspired Unmanned Aerial Vehicle (UAV). Unlike majority of the current popular configurations of UAVs such as multirotors and fixed-wing aircraft, this particular nature-inspired UAV generates lift by spinning its entire body about a central axis. This inherent oscillatory nature makes this particular type of UAV suitable for the implementation of a CPG-based controller, which also possess oscillatory characteristics. The CPG handles the low-level actuator commands given a simple higher-level input. Policy Gradients with Parameter-based Exploration (PGPE) was used to optimize the CPG parameters to obtain the desired UAV motion from the CPG inputs. A Tip Path Plane (TPP) angle controller was added on top of the CPG to form a TPP-CPG controller, in which closed-loop position control was implemented above this. Hovering and trajectory tracking tests were successfully conducted and the performance of the control strategy was verified.