Investigation on a bipedal robot: Why do humans need both Soleus and Gastrocnemius muscles for ankle push-off during walking? Bernadett Kiss 1 , Emre Cemal Gonen 1 , An Mo 1 , Alexandra Buchmann 2 , Daniel Renjewski 2 and Alexander Badri-Spröwitz 1 Abstract— Legged locomotion in humans is influenced by mechanics and neural control. One mechanism assumed to contribute to the high efficiency of human walking is the impulsive ankle push-off, which potentially powers the human swing leg catapult. However, the mechanics of the human’s lower leg with its complex muscle-tendon units spanning over single and multiple joints is not yet understood. Legged robots allow testing the interaction between complex leg mechanics, control, and environment in real-world walking gait. We custom developed a small, 2.2 kg human-like bipedal robot with soleus and gastrocnemius muscles represented by linear springs, acting as mono- and biarticular elasticities around the robot’s ankle and knee joints. We tested the influence of three soleus and gastrocnemius spring configurations on the ankle power curves, on the synchronization of the ankle and knee joint movements, on the total cost of transport, and on walking speed. We controlled the robot with a feed-forward central pattern generator, leading to walking speeds between 0.35 m/s and 0.57 m/s at 1.0 Hz locomotion frequency, at 0.35 m leg length. We found differences between all three configurations; the soleus spring supports the robot’s speed and energy efficiency by ankle power amplification, while the GAS spring facilitates the synchronization between knee and ankle joints during push-off. I. INTRODUCTION How complex the human bipedal walking is becomes apparent when attempting to technically replicate or restore the human lower limb’s musculoskeletal system, e.g., with humanoid robots or lower limb prostheses. Robots and prostheses do not yet reach human performance in terms of efficiency, mobility, and robustness. The current gap implies a missing understanding of biomechanics and control of human locomotion. One mechanism assumed to contribute to the high efficiency of human walking is the impulsive ankle push-off, which potentially powers the human swing leg catapult (SLC) [1]. A catapult’s function is physically characterized by slow storage of elastic energy, followed by a rapid release of stored energy, with a substantially higher output power that accelerates a projectile. Hof et al. described first how the ankle power burst in the late stance phase is preceded by a slower energy storage phase *Gefördert durch die Deutsche Forschungsgemeinschaft (DFG)- 449427815, 449912641, and 3991/2-1, and supported by the International Max Planck Research School for Intelligent Systems, the China Scholarship Council and the Max Planck Society. 1 Dynamic Locomotion Group, Max Planck Institute for Intelligent Systems, 70569 Stuttgart, Germany [kiss;gonen;mo;sprowitz]@is.mpg.de 2 Chair of Applied Mechanics, TUM School of Engineering & Design, Department of Mechanical Engineering, Technical University of Munich, 85748 Garching near Munich, Germany [firstname.lastname]@tum.de in human walking [2]. The observed ankle power burst is hereby higher than the plantar flexor muscles’ peak power [1], [3], [4]. This indicates that additional passive structures store elastic energy as kinetic energy [2], [5]. The rapid release of the stored elastic energy at the end of stance is available to accelerate the stance leg into swing like a catapult firing its projectile—the swing leg [1]. A catapult has three main mechanical components: an elastic element, a block, and a catch with or without escapement. In the human lower leg, the complex interplay between the thigh- shank-foot segment chain and muscle-tendon-units (MTU) spanning the ankle joint make it hard to identify the exact swing leg catapult mechanics and its functionality. Elastic energy storage in stance and rapid recoil during push-off is facilitated by the soleus (SOL) and gastrocnemius (GAS) muscles [3], [4], [6], [7]. The GAS muscle works mostly isometrically during stance, and potentially facilitates energy efficient loading of its muscle tendon unit. In contrast, the SOL muscle seems to contribute to ankle coordination by active contraction [8]. Up to 91 % of the power output during push-off is provided by elastic energy [9]. Changes in the effective ankle stiffness influence the energy efficiency of walking, as studies with passive ankle foot orthosis showed [10], [11]. The combination of ground and foot is the human leg catapult’s block, against which the swing leg catapult unloads when shifting the body weight from the stance to the trailing leg. This happens in the very brief double support phase of walking, during late stance. Interestingly, the catch of the swing leg catapult has not been identified in the human leg yet. Previous investigations into multi-articulate actuators and spring-tendons in robots showed several advantages. The BioBiped3 robot with biarticular muscles showed improved balance control during upright standing and locomotion, and axial leg function during bouncing [12]. CARL robot improved its hopping efficiency by 16 % by using mul- tiarticulate actuators [13]. A simulation study shows that combining mono- and bi-articular foot prosthesis actuation reduces peak power requirements [14]. So far, the exact role and function of plantarflexor spring-tendons (GAS and SOL) as part of the swing leg catapult during walking has not been studied in robots. In this work we show the influence of SOL and GAS muscles on a human-like bipedal robot’s ankle power curve, ankle and knee joint synchronization, energy efficiency and walking speed. As testing platform the developed robot features both ankle muscles as linear springs. We analyze the robot’s energy consumption for one gait cycle and report the arXiv:2203.01588v1 [cs.RO] 3 Mar 2022