Content area
Full Text
1.
Introduction
The cerebellum controls and plans the balance and locomotion of the animals. Moreover, the vestibular system and the proprioception system acting as biological sensors also play key roles in stabilizing the body posture and planning the movement (Deliagina et al. , 2007). The vestibular system could detect the angular displacement of body in three dimensions. In addition, with the capability of sensing the gravitational acceleration, the vestibular system could act as a feedback sensor for detecting the body attitude (Lopez and Blanke, 2011). Meanwhile, the proprioception system acts as another important feedback sensor. The angle and angular velocity of the joint, the tension of the muscle and the location of the body segments could be perceived by the proprioception system (van Beers et al. , 1998).
The spinal cord of an animal's vertebrate has neural circuits called central pattern generators (CPGs), which could produce rhythmic patterns of neural activity (Ijspeert et al. , 2007). Many fundamental rhythmic activities are actuated by the CPGs, such as chewing, breathing and digesting. The stable gait could also be regarded as a cyclic motion. As a result, the CPG is suitable for producing the rhythms of the leg movements in stable quadruped locomotion. Spröwitz et al. presented a new CPG model which consisted of a fully connected network of four phase-coupled oscillators (Sprowitz et al. , 2013), and each oscillator drove one hip joint. All the control parameters of the CPG were optimized with the particle swarm optimization (PSO) algorithm. With the CPG model, cheetah-cub realized a stable trotting both in simulation and experiment. He et al. (2015) used the proprioception system to capture the actual output joint angles produced by the CPG. And then, the achieved data were transferred to the cerebellum for comparison with the parameter values from the cerebral cortex. Finally, the calculated compensation was sent to adjust the CPG network. With the bio-inspired control strategy, a human walking gait was realized based on the seven-link biped robot model. Fukuoka et al. (2013) built a planar model equipped with a CPG which consisted of four coupled neural oscillators. In addition, the vestibular feedback was introduced in the CPG model to control the pitching motion of the trunk. Moreover, the proprioceptive sensory feedback was...