The paper reports on a project to make a quadruped robot walk with medium forward speed on irregular terrain in an outdoor environment using a neural system model. The KEY WORDS-legged locomotion, quadruped, biologically inspired robot, adaptive walking on irregular terrain, neural system model, central pattern generator (CPG), reflex

We have been trying to induce a quadruped robot to walk with medium walking speed on irregular terrain based on biological concepts. We propose the necessary conditions for stable dynamic walking on irregular terrain in general, and we design the mechanical system and the neural system by comparing biological concepts with those necessary conditions described in physical terms. A PD controller at the joints can construct the virtual spring-damper system as the visco-elasticity model of a muscle. The neural system model consists of a central pattern generator (CPG) and reflexes. A CPG receives sensory input and changes the period of its own active phase. The desired angle and P-gain of each joint in the virtual springdamper system is switched based on the phase signal of the CPG. CPGs, the motion of the virtual spring-damper system of each leg and the rolling motion of the body are mutually entrained through the rolling motion feedback to CPGs, and can generate adaptive walking. We report on our experimental results of dynamic walking on terrains of medium degrees of irregularity in order to verify the effectiveness of the designed neuro-mechanical system. We point out the trade-off problem between the stability and the energy consumption in determining the cyclic period of walking on irregular terrain, and we show one example to solve this problem. MPEG footage of these experiments can be seen at http://www.kimura.is.uec.ac.jp.

We have been trying to induce a quadruped robot to walk with medium walking speed on irregular terrain based on biological concepts. We propose the necessary conditions for stable dynamic walking on irregular terrain in general, and we design the mechanical system and the neural system by comparing biological concepts with those necessary conditions described in physical terms. A PD controller at the joints can construct the virtual spring-damper system as the visco-elasticity model of a muscle. The neural system model consists of a central pattern generator (CPG) and reflexes. A CPG receives sensory input and changes the period of its own active phase. The desired angle and P-gain of each joint in the virtual springdamper system is switched based on the phase signal of the CPG. CPGs, the motion of the virtual spring-damper system of each leg and the rolling motion of the body are mutually entrained through the rolling motion feedback to CPGs, and can generate adaptive walking. We report on our experimental results of dynamic walking on terrains of medium degrees of irregularity in order to verify the effectiveness of the designed neuro-mechanical system. We point out the trade-off problem between the stability and the energy consumption in determining the cyclic period of walking on irregular terrain, and we show one example to solve this problem. MPEG footage of these experiments can be seen at http://www.kimura.is.uec.ac.jp.

We discovered a specific rule for generating typical quadrupedal gaits (the order of the movement of four legs) through a simulated quadrupedal locomotion, in which unprogrammed gaits (diagonal/lateral sequence walks, left/right-lead canters, and left/right-lead transverse gallops) spontaneously emerged because of leg loading feedbacks to the CPGs hard-wired to produce a default trot. Additionally, all gaits transitioned according to speed, as seen in animals. We have therefore hypothesized that various gaits derive from a trot because of posture control through leg loading feedback. The body tilt on the two support legs of each diagonal pair during trotting was classified into three types (level, tilted up, or tilted down) according to speed. The load difference between the two legs led to the phase difference between their CPGs via the loading feedbacks, resulting in nine gaits (32: three tilts to the power of two diagonal pairs) including the aforementioned.