Optimum gait for quadruped robot using multi-objective Jaya optimization algorithm

Creating a four-legged robot's gait means creating movements in the robot's legs to ensure the robot walks naturally and saves energy. Currently, this is still a difficult problem because current technology cannot access four-legged animals with extremely complex structures and sophisticated movements. This article proposes an optimal gait generation model for robot dogs that examines the trade-off between stability and speed. Firstly, the trajectory of the hips and feet at each leg depends on four gait parameters (step length, leg lift, knee bend, and hip swing) that will be set based on the movement rules of the robot dog and the third-order interpolation function. Second, 12 joint-angle orbits at the four legs of the robot dog will be deduced from the orbits of the hips and the orbits of the feet at the four legs of the robot dog by solving the inverse kinematics problem using analytical method. Then, a multi-objective function is built concerning speed and stability, based on the gait characteristics (gait parameters, CoP/ZMP trajectory) of the robot dog. Eventually, the Jaya multi-objective optimization algorithm (MO-JAYA) is applied to find four optimum gait parameters, so that the robot dog performs a stable walk at the fastest possible speed. This proposed method is tested on the small-sized quadruped robot (B3-BOT). The simulation results demonstrate that B3-BOT quadruped robot walks steadily at the fastest possible speed.