CPG-based control design for bipedal walking on unknown slope surfaces

The paper presents a walking pattern generator and a balance control system for a bipedal robot to handle an unknown slope. The robot uses onboard gyro and accelerometer sensors to detect the pose information of the upper-body. A controller is proposed for the robot to walk on an unknown slope by adjusting the tilt angle of the upper-body. The theory of central pattern generator (CPG) is applied to generate the walking trajectory. By using the pose information of the upper-body, we developed a method to determine the relationship between the slope surface and the upper-body pose and generate the compensation motion to adjust the tilt angle of the upper-body. The compensation control consists of predictive compensation and immediate compensation. The predictive compensation responds to adjust the upper-body pose before beginning of the next step. The immediate compensation is applied to adjust the upper-body pose during the single support phase. The integrated controller adapts to the unknown slope in real time while robot walking. Using the bipedal robot NAO, the experimental results show that the biped robot can walk successfully on unknown slopes.