Design of the musculoskeletal leg $\rm C \tiny{\rm ARL}$ based on the physiology of mono-articular and biarticular muscles in the human leg

Abstract

In a lower extremity musculoskeletal leg, the actuation kinematics define the interaction of the actuators with each other and the environment. Design of such a kinematic chain is challenging due to the existence of the redundant biarticular actuators which simultaneously act on two joints, generating a parallel mechanism. Actuator kinematics is mainly dependent on the moment arm profile of the actuation system. It is a common practice to select a constant moment arm value for robotic actuation system; nevertheless, biological muscles feature a distinctive nonlinear moment arm profile that has been ignored in the design of the musculoskeletal robots. In this paper, we propose a design paradigm for compliant robotic leg Carl based on the direct replication of the human leg anatomy. The resulting mechanical system should (a) demonstrate a similar moment arm profile as in leg musculature, (b) exhibit expected physiological behavior of the muscles, (c) provide insight into the interaction of the actuators and possible improvement in the efficiency of the movements. We provide a comprehensive analysis of the moment arm profile of the leg musculature. The actuator kinematics of the designed leg is validated by comparing the contraction velocities of the muscles and actuators. The biological characteristics of the actuators are analyzed using the jump experiment data conducted on the previous version of the leg. The major physiological characteristics of the biarticular muscles, ligamentous action, and distal power transfer, is successfully demonstrated by the robotic leg. Our analysis demonstrates that the proposed structural design of the actuation system can improve the mechanical efficiency of this particular jump experiment up to 16% compared to the leg without actuator redundancy. Compared to the previous version of the leg, by only modifying the moment arm profiles, we can achieve an efficiency improvement of approximately 5%.

Technical advantages and disadvantages of biarticular actuators in bipedal robots

“…As a result, the mechanical efficiency of the jump movement is improved by 18%. 10 Moreover, During explosive leg extension of the robotic jump experiment, the maximum contraction velocity of biarticular A H K is 0.1 m s −1 which is 4 times less than contraction velocity of mono-articular at knee A K and 2 times less than hip mono-articular A H , suggesting the bio-inspired ligamentous action of BiAs. These observations demonstrate that the efficiency achieved by the addition of biarticular actuators can indeed justify the extra added weight of redundant actuators in explosive movements.…”

“…In the human leg, the upper limb pseudo-antagonist biarticular muscles RF and HAM exhibit during various locomotion type an average hip-to-knee moment arm ratio of 0.7 and 2.0, respectively. 10 Carl is designed to have one upper limb biarticular A H K acting as a compromise between function of RF and HAM with RMA of 1.0. 10 The lower limb biarticular in humans is GA with an average ankle-to-knee moment arm ratio of 2.0, the similar ratio is adopted in the design of biarticular actuator A K A in Carl.…”

“…10 Carl is designed to have one upper limb biarticular A H K acting as a compromise between function of RF and HAM with RMA of 1.0. 10 The lower limb biarticular in humans is GA with an average ankle-to-knee moment arm ratio of 2.0, the similar ratio is adopted in the design of biarticular actuator A K A in Carl. The moment arm of the BiAs are depicted in Fig.…”

“…The nonlinear MAP for BiAs is selected such that it exhibits the corresponding RMA during the joint trajectory of human walking, running and hopping. 10…”

Diaphragm Ankle Actuation for Efficient Series Elastic Legged Robot Hopping

“…Control and design complexity increase when actuation is transferred over multiple joints, as is the case with robotic legs that feature more than two actuated segments. Multijoint transmission can be achieved with cable (tendon), chain, and belt systems [3], [4], [5]. Simple configurations lead to mechanical coupling, where the output of one motor affects the movement of several, in-between connected joints.…”