A Novel Nonlinear CPG-Based Oscillator Model for Quadruped Robotic Locomotion

In recent decades, robotic locomotion has applied different techniques to emulate central pattern generators (CPGs). The theory of CPG explains the biological functions of motor control in living organisms. This paper presents an unpublished model for coupled nonlinear oscillators. This model employs a canonical nonlinear differential equation system to coordinate joint activity. The analysis, conducted under the criteria of chaos and bifurcation theory, determines that the new model is successful and without the presence of chaos. The new model is compared with other cases, including the Wilson–Cowan, Hopf, and Van Der Pol models, as well as with the operability of different robots. It highlights the new model’s advantages in terms of versatility, simplicity, and processing, as well as the comparisons of metrics of locomotion, such as support factor and symmetry index between hemibody metrics. The new model is applied to the locomotion of two quadruped robots (a crab and a dog) used in research on transitions between types of locomotion, considering both physical and computational limitations.