An in silico central pattern generator: silicon oscillator, coupling, entrainment, and physical computation

Abstract

In biological systems, the task of computing a gait trajectory is shared between the biomechanical and nervous systems. We take the perspective that both of these seemingly different computations are examples of physical computation. Here we describe the progress that has been made toward building a minimal biped system that illustrates this idea. We embed a significant portion of the computation in physical devices, such as capacitors and transistors, to underline the potential power of emphasizing the understanding of physical computation. We describe results in the exploitation of physical computation by (1) using a passive knee to assist in dynamics computation, (2) using an oscillator to drive a monoped mechanism based on the passive knee, (3) using sensory entrainment to coordinate the mechanics with the neural oscillator, (4) coupling two such systems together mechanically at the hip and computationally via the resulting two oscillators to create a biped mechanism, and (5) demonstrating the resulting gait generation in the biped mechanism.

Biologically inspired adaptive walking of a quadruped robot

“…A great deal of the previous research on this attempted to generate walking using a neural system model, including studies on dynamic walking in simulation (Taga et al 1991;Ijspeert 2001;Tomita & Yano 2003) and real robots (Ilg et al 1999;Kimura et al 1999;Tsujita et al 2001;Lewis et al 2003). But autonomously adaptive dynamic walking on irregular terrain was rarely realized in those earlier studies except for our studies using quadruped robots called 'Patrush' (Kimura et al 1999(Kimura et al , 2001) and 'Tekken1' (Fukuoka et al 2003).…”

Spinal Information Processing and its Application to Motor Learning Support

“…Studies focusing on the neuromuscular system are well known in walking robot control [13][14][15][16][17] and estimation of arm posture from surface EMG signals [18,19]. We have proposed extracting motor features based on the structure of the human neuromuscular system by dividing the whole-body motion pattern into basic patterns by nerve.…”

Evolution of central pattern generators for the control of a five-link bipedal walking mechanism

“…CPGs have also been successfully applied to non-legged cases such as serpentine locomotion [26] and swimming [27]. It has also been demonstrated that CPG controllers can be implemented as analog electronic circuits [20]. Important advantages of legged robotic locomotion with CPGs are: being biologically inspired and suited for distributed implementation, having few control variables, and exhibiting stable limit cycle behavior resistant to perturbations, as compared to classical control approaches such as the widely used zero moment point (ZMP) method, finite-state machines, or heuristic control [14].…”

“…This has the advantage of making the physical simulation and mathematical analysis of the gait less complicated, providing insight on some of the central processes that occur in the lateral plane during bipedal locomotion. Mechanisms of this type have been used on many occasions to study bipedal walking [17,18,20,23,30] and running [32]. The fivelink mechanism that we use in this study as a physics simulation and constructed hardware is comparable with the one used by Lewis et al [18], and also by Geng, Porr, and Wörgötter [13] and Pratt et al [19], in terms of general structure and the arrangement of lateral support.…”

“…We apply CPG control to a five-link planar bipedal walking mechanism [11,18], a minimalistic structure containing four actuators (corresponding to the hip and knee of each leg) and touching the ground with rounded extremities without ankles ( Figure 6). Presented mathematical model and simulation are based on the work by Taga et al [11]; and by using a minimal set of actuators and a planar support structure, we aim to restrict this current study to the smallest set of parameters and two dimensions, comparable to several existing studies of bipedal walking [12,19,20]. We also present a test of the best evolved network on a physical five-link walking mechanism, to investigate whether CPG networks evolved using simulations can be acceptably transferred to real robot hardware.…”