Research for gait stability of biped walking using vision

The paper is about a biped walking robot that has 12 degrees of freedom and achieves proper gate stability with visual feedback during walking. The bipedal robot is equipped with cheap FSR sensors those are attached underneath the soles of the robot to obtain ZMP instantaneously. It uses a cheap USB camera that is attaching on its head to process images of environment during walking. With the processed images the robot can adjust its posture and its gait trajectory to increase stability based on ZMP and tilting algorithm. The robot extracts a possible line feature from the image of the surrounding environment with dynamic contour-based method for fast detection. Once the target line segment is recognized, the orientation angle and the position of that line feature are continuously monitored. If those are deviated more than certain threshold, tilting its body and adjusting the walking trajectory are executed to keeps the biped robot walking stable.