Generation of Inverted Locomotion Gait for Multi-Legged Robots Using a Spherical Magnet Joint and Adjustable Sleeve

Abstract

In this paper, we propose a design and an implementation of spherical magnet joint (SMJ) - based gait generation for inverted locomotion of multi-legged robots. A spherical permanent magnet is selected to generate a consistent attractive force for the robot to perform inverted locomotion under steel structures. Additionally, the tip of the robot's foot is designed as a ball-joint mechanism to give flexibility to the foot placement at any angle between the tip and surfaces. We also propose an adjustable sleeve mechanism to detach the tip of the foot during locomotion by creating a fulcrum point during the tilt and pull step. As a result, the reaction force can be reduced according to sleeve diameter. Experimental results show that the presented load decreased by 46% from direct pulling with the adjustable sleeve mechanism. For inverted locomotion, a quadruped robot and a hexapod robot were constructed to represent the predominant type of multi-legged robot. We integrated the SMJ and the adjustable sleeve on both robots and performed the inverted locomotion with a crawling gait, a trotting gait, a square gait, and a tripod gait. Our analysis demonstrates the characteristics of each gait in terms of velocity, stability, guaranteeing the versatility of our proposed SMJ, which can be applied to different types of legged robots.

Anthropomorphic walking robot: Design and simulation

“…Unlike the technical solutions of walking robots discussed above [2,3,6,13], which have a separate drive for each leg joint, the proposed anthropomorphic walking mechanism contains a minimum of drives for each leg. In relation to other structures considered in Section 2, the proposed walking mechanism has a significantly lower cost due to the minimization of drives and the use of sealed gas chambers for energy recovery.…”

“…The analysis of models of anthropomorphic robots presented in [12] illustrates the possibilities of machine learning for walking robots based on the algorithms of bionic structures. In [13], the designs of walking robots with four and six legs are proposed. However, in these works, there are no recommendations for improving the energy efficiency of walking robots by reducing the drives.…”

“…Next, the total time derivatives are calculated, and the possible increments of the movement coordinates and rotation angles of the robot's legs are specified. As a result, we obtain the formula for the generalized force: (13) where: M 2 -is the torque of the electric motor; g=9.81 m/s 2 -acceleration of free fall, see other designations above.…”