Stable Locomotion of Humanoid Robots on Uneven Terrain employing Enhanced DAYANI Arc Contour Intelligent Algorithm

Abstract

Humanoid robots must be capable of walking on complicated terrains and tackling a variety of obstacles leading to their wide range of possible implementations. To that aim, in this article, the issue of humanoid robots walking on uneven terrain and tackling static and dynamic obstacles is examined. It is inspected by implementing a novel Enhanced DAYANI Arc Contour Intelligent (EDACI) Algorithm that designs trajectory by searching feasible points in the environment. It provides an optimum steering angle, and step optimization is performed by BFGS (Broyden–Fletcher–Goldfarb–Shanno) Quasi-Newton method that leads to guide the humanoid robot stably to the target. The leg length policy has been presented and a reward-based system has been implemented in the walking pattern generator that provides the optimum gait parameters. One humanoid robot act as a dynamic obstacle to others if they are navigating on a single terrain. It may generate a situation of deadlock, which needs to be solved. In this article, a dining philosopher controller (DPC) is employed to deal with and solve this issue. Simulations are used to evaluate the proposed approach in several uneven terrains having two humanoid NAOs. The findings indicate that it can precisely and efficiently produce optimal collision-free paths, demonstrating its efficacy. Experiments in similar terrain are performed that validate the results with a deviation under 6 %. The energy efficiency of the developed controller is evaluated in reference to the inbuilt controller of NAO based on energy consumption. In order to check the feasibility and accuracy of the developed controller, a comparison with an established technique is provided.