Research And Implementation On Motion Control Method Of Quadruped Robot On Complex Terrain And Environment

More than half of the Earth’s land surface are rugged terrains, such as mountain, swamp and ravine. Traditional wheeled or tracked vehicles can not move on these types of terrains, so tasks such as geographical terrain surveying, engineering exploration and military reconnaissance can not be completed using traditional mobile robots. However, many terrestrial animals can run and jump flexibly on these types of terrains, which give people some inspiration, legged moving method has great advantage on rugged terrain.There are millions of creatures in nature, while most of large terrestrial animals are quadrupeds, which indicate that quadrupeds gained the recognition of nature in the long period of brutal Darwinian evolution, and gradually grew and multipliedTto date. Cheetahs can run at high speed, goats can walk on the cliffs, and camels can walk with heavy load in harsh environments for a long time. The great advantage of four-legged moving mode is self-evident.So far, legged robots can be devided into several types according to the number of legs, such as one leg, two legs, four legs, six legs and eight legs. Quadruped robots have greater stability than biped robots, and the structure are more simple than hexapod robots and octopod robots, so much attention is paid on them, which is entirely consistent with the choice of nature, and the determination of studying quadruped robots is confirmed.By studying the structure and gaits of quadruped robots, a series of control methods are proposed, the content mainly includes the following aspects:1. For mobile robots, moving is the basic function of the robot. For quadruped robots, fast and flexible moving while keeping balance is of great importance in the control of the robot. In this paper, a motion control method based on overlaying oscillations is proposed. First, the motion of the robot in plane is decoupled into three parts, the forward and backward motion, the leftward and rightward motion, and the rotary motion. Then, the speed of the three directions are controlled independently, after that, overlay the trajectories to get the final toe trajectory, which implemented the omni-directional moving of the robot.2. Slope is a common terrain, when the robot moves on a slope, the projection of the center of mass of the robot in the support polygon will shift downward the slope, which affects the symmetry of the robot, reduces the stability of the motion. In this paper, the kinematics of the robot is analyzed, by mapping the coordinate systems, the position of the toes are modified, the robot adepts to the slope without affecting the gait planning part, and the omni-directional moving method is expended to various of slopes, just like moving on flat terrain.3. Except for level ground and slopes, there are various of rugged terrains, how to move on unstructured rugged terrain stably is the key and difficult part of quadruped motion control. In this paper, tactile is used to detect the unstructured terrain in real time, and the motion of the robot is decoupled into basic gait planning, adjustment of the position of center of mass and adjustment of posture, the stable motion of the robot is implemented by coordinating these tasks. Finally, the robot is tested on a broken staircase. The robot climbs up the stairs successfully without any prior knowledge of the environment, which shows a good stability and locomotivity on rugged terrain.4. The sections above are about robot motion control by planning the positions of the toes, in addition, the balance control problem of the robot standing with diagonal legs in force control mode is also studied. By now most legged robots keep stable relying on large sole of the feet, more legs or continuously jumping, however, in this paper, the control of the robot standing with diagonal legs is decoupled into three parts and these three parts are controlled independently, methods such as linear quadratic regulator and virtual model control are used to keep the stability of the parts, and the robot achieved a stable state of standing on two legs, while the toes and ground are point contact and the robot need not to keep jumping. In simulation, the robot kept stable under posture disturbance, torsional disturbance and even lateral disturbance, which proved the effectiveness and strong stability of the control method.