| The Platform for Ambulating Wheels (PAW) is a hybrid quadruped wheeled-legged robot that can bound, gallop, roll and brake at high speeds, and perform inclined turning. In previous work, the PAW's controller used fixed touchdown and liftoff angles to achieve a stable bounding gait, and these angles were predetermined through an extensive trial and error process. In this work, an intelligent velocity controller is developed to allow the robot to autonomously find the touchdown and liftoff angles to bound at a desired velocity. This enables the robot to track desired velocities between 0.9 and 1.3 m/s, as shown in a Matlab-Adams co-simulation model of bounding. The controller also demonstrates tracking capabilities in the presence of minor terrain changes.Finally, the intelligent controller is implemented and tested on the physical platform demonstrating adequate velocity tracking for set points between 0.9 m/s and 1.3 m/s, as well as transitions between set points in this range.To implement this controller on the physical platform, an Extended Kalman Filter (EKF) is developed to estimate the forward velocity of the robot required as a controller input. The EKF combines the data from an Inertial Measurement Unit and an estimate of forward velocity found kinematically using measurements from motor encoders and leg potentiometers. The accuracy of the EKF estimate of the forward velocity is validated in simulation and using high speed camera experiments. |
