Leg design for biped locomotion with mono-articular and bi-articular linear actuation. (February 2021)

• Record Type: Journal Article
• Authors: Chevallereau, Christine; Wenger, Philippe; Aoustin, Yannick; Mercier, Franck; Delanoue, Nicolas; Lucidarme, Philippe
• Is Part Of: Mechanism and machine theory. Volume 156(2021)
• Publication Date: 2021-02
• Subjects: Biped leg -- Optimal design -- Linear motor -- Mono-articular -- Bi-articular -- Energy; Machine theory -- Periodicals; Machinery -- Periodicals; Machines -- Périodiques; Génie mécanique -- Périodiques; Machine theory; Machinery; Periodicals; 621.81
• DOI: 10.1016/j.mechmachtheory.2020.104138
• Languages: English
• ISSNs: 0094-114X

Abstract

Highlights: An original design strategy of biped legs actuated with linear motors is proposed. Eight actuation schemes combining mono-articular and bi-articular motors are optimized. The best actuation scheme with three motors and their optimal attachment points are provided. A first analytic-heuristic approach is provided to obtain more insight into the design. A numerical optimization is run with the starting solutions obtained from the analytic step. Abstract: Implementing efficient actuators with good backdriveability properties is of primary importance for the design of high-performance performance bipedal robots. Nowadays, the use of linear direct drive electric motors is an interesting choice. This actuation scheme is closer to the efficient muscular system of human legs. However, the efficiency of linear motors is highly sensitive to the placement of their attachment points. Two main issues are addressed in this paper; (i) the optimal placement of the motors on the leg and (ii) the best actuation scheme of the leg with three motors when considering both mono-articular and bi-articular motors. Little research work has been devoted to these issues. Accordingly, a methodology is proposed to optimize linear motors attachment points. Instead of resorting to a pure numerical optimization, a first analytic-heuristic approach is proposed in order to obtain more insight into the optimization problem. This first approach, in addition to providing useful information, allows obtaining a good initial guess for the numerical optimization. Eventually, eight different leg architectures with three motors and combining different numbers of mono-articular and bi-articular motors are optimized. Results show the interest of using bi-articular motors to reduce the maximum effort required for each motor.