Parameter Tuning of ODE-Based Walking Control Algorithm for a Six Legged Mobile Robot

Abstract

The design of controllers which considering the dynamic effects that walking robots may encounter during their movement, is an important requirement in the field of legged mobile robots. In recent years, in the control of complex dynamical systems, numerical simulation algorithms based control structures, have begun to be preferred. Open Dynamics Engine (ODE) is a physics simulator for modeling and controlling of rigid bodies. In the ODE-based stabilization and walking algorithm that earlier developed for a six-legged mobile robot, a model based controller structure was obtained that takes into account the disturbance effects that occur during control. In this study, the parameters of ODE-based walking and stabilization algorithm were tuned. It is known that the sequential sampling strategy gives affirmative results in tuning the parameters of model-based controllers. For this reason, a sequential sampling strategy was used to tune the parameters of the ODE-based stabilization and walking controller. After this tuning procedure, 37.83% improvement in area error and 4.63% improvement in similarity error were obtained.