Three-dimensional walking of a simulated muscle-driven quadruped robot with neuromorphic two-level central pattern generators

Abstract

We aim to design a neuromorphic controller for the locomotion of a quadruped robot with muscle-driven leg mechanisms. To this end, we use a simulated cat model; each leg of the model is equipped with three joints driven by six muscle models incorporating two-joint muscles. For each leg, we use a two-level central pattern generator consisting of a rhythm generation part to produce basic rhythms and a pattern formation part to synergistically activate a different set of muscles in each of the four sequential phases (swing, touchdown, stance, and liftoff). Conventionally, it was difficult for a quadruped model with such realistic neural systems and muscle-driven leg mechanisms to walk even on flat terrain, but because of our improved neural and mechanical components, our quadruped model succeeds in reproducing motoneuron activations and leg trajectories similar to those in cats and achieves stable three-dimensional locomotion at a variety of speeds. Moreover, the quadruped is capable of walking upslope and over irregular terrains and adapting to perturbations, even without adjusting the parameters.

Autonomous speed adaptation by a muscle-driven hind leg robot modeled on a cat without intervention from brain

“…Although our CPG model is simpler than the comparable Markin et al's CPG model 16 (as described in the "Introduction" section), we consider ours practical enough for use as a biomimetic locomotion controller for a robot, due to its autonomous speed adaptation ability as demonstrated in this article, as well as stable quadrupedal walking shown in a previous simulation. 9…”

“…We briefly review the mechanism of each hind leg, which is the same as in our previous cat-simulated model. 9 Each hind leg has the hip, knee, and ankle joints around the pitch axis, and these joints are driven by six representative muscles, including two two-joint muscles. The muscle placement in each hind leg is shown in Figure 1(b).…”

“…We used our previously proposed two-level CPG model 9 for our present cat hind leg simulated model and robot with the exception of the Inab neuron to implement a stretch reflex. Since the stretch reflex had mainly worked to ensure walking stability in the lateral plane in our previous study, 9 it was not needed for planar locomotion in this article. The equations of each neuron model used in the two-level CPG model (equations (2)-( 5)) are briefly reviewed in the "Basic design of two-level CPG model" section.…”

“…We proposed a two-level CPG model to drive a single leg with three joints in our previous simulation work. 9 The CPG consisted of a rhythm generation (RG) part to generate basic locomotion rhythms and a pattern formation (PF) part to synergistically activate a different set of muscles in each of the four sequential phases (swing, touchdown, stance, and liftoff). There were motoneurons (Mn) under the PF part to activate the corresponding muscle according to the driving signal from the PF neurons.…”

Learning Free Gait Transition for Quadruped Robots Via Phase-Guided Controller

“…The balance can be achieved with manually crafted feedback. Those works mainly focus on the structure of CPG network [17], the mathematical forms [14], [15], [18], the design of feedback control [19], [20] and the ability to transit among various gaits [13], [21]. Unfortunately, none of the above works have demonstrated agile and robust locomotion with quadruped robots.…”