Aperiodic walking control of a biped robot perturbed by continuous strong external forces

When robots walk on rough terrain or are subjected to continuous strong external disturbances, traditional methods do not have the flexibility to adjust bipedal walking gaits in a timely manner, often resulting in walking instability. In response to this issue, an angular momentum linear inverted pendulum model and a multi-link dynamic model were established, respectively, for the biped robot Rabbit walking on an inclined terrain and disturbed by external forces at its center of mass. A controllable domain for the biped robot was defined by choosing the single-step average speed as the controllable target, from which the influence of variations of single-step duration and length was briefly analyzed. Based on the single-step duration optimization model combined with the traditional single-step length optimization scheme, a dual parameter optimization algorithm was developed. Subsequently, by designing a hierarchical control strategy based on the optimized gait parameters and joint trajectory tracking, an adaptive aperiodic gait of the biped robot was ultimately achieved. The simulation results of Rabbit walking on flat and inclined terrains under four different continuous strong external forces show that using the traditional single-step optimization scheme will lead to walking instability within a very limited time. However, a hierarchical anti-disturbance control strategy based on dual parameter optimization in the centroid layer and trajectory tracking in the joint layer can enable Rabbit to continue walking without falling for a longer time. In addition, the dual parameter optimization method can not only expand the controllable domain, but also effectively adjust the average speed of each step of Rabbit, making it change as much as possible within a safe and controllable range, and generating a more flexible and natural aperiodic gait.