Bond Graph Modeling and Optimization of a Bipedal Robot Leg for Enhanced Walking Stability

Compared to wheeled robots, legged robots have more instability control because their leg lifting and placing causes irregular ground contact. As the field of robotics evolves, the use of Bond Graph Modeling in the design optimization of bipedal locomotion systems offers a promising avenue for advancing humanoid robot technology and achieving more natural and efficient movement patterns. The objective of this paper is to model and simulate an enhanced walking mechanism for a bipedal robot using the bond graph model for the bipedal robot, integrating the use of optimization technique, namely sweeping parameters of the robot's coefficients to enhance and control the robot's locomotion, to support the results of the simulations that enhances and stabilizes the robot gait cycle using 20-Sim Software. The results are promising compared to the practical actual real robot which show the power of bond graph modeling which define the robot by its elements to perform smooth gait cycles. This study paves the way for future research on controlling humanoid robots by utilizing an artificial neural network controller combined with a bond graph model to improve the gait cycle.