For almost seven decades, inventors and researchers have dedicated themselves to the development of robots. Until now, all machines fabricated and deployed by humans, regardless of their application, share a common thread: they are driven by electric motors, a technology that has been stagnant for over two centuries. While strolling robots do feature legs and arms, these are actually driven by motors rather than relying on muscle tissue like humans and animals do? As a result, they appear to lack the mobility and adaptability characteristic of living organisms.
A cutting-edge muscle-powered robotic leg boasts not only enhanced energy efficiency but also enables high-jumping capabilities and rapid movements, all while detecting and responding to obstacles without reliance on complex sensors. Researchers at ETH Zurich and the Max Planck Institute for Intelligent Systems have collaborated on the development of a novel leg, within the framework of the Max Planck ETH Center for Learning Systems (CLS). The CLS workforce was spearheaded by Dr. Robert Katzschmann at ETH Zurich and Prof. Christoph Keplinger at the Max Planck Institute for Intelligent Systems (MPI-IS). Thomas Buchner and Toshihiko Fukushima, co-first authors alongside their doctoral colleagues, published a groundbreaking study on an animal-inspired musculoskeletal robotic leg in Nature Communications.
An analogy exists between the musculature of living beings and the control systems of robots, allowing for smooth transitions between extension and flexion in robotic legs. Researchers have developed innovative electro-hydraulic actuators, dubbed HASELs, that are connected to skeletons via tendon-like structures.
The actuators are oil-filled plastic tubes resembling those used in traditional ice cube trays. Approximately half of each bag’s surface area is treated with a uniform coating of a conductive material-based black electrode applied to both sides. As soon as a voltage is applied to the electrodes, they rapidly attract each other due to the instantaneous discharge of static electrical energy. As I move the balloon close to my scalp, my hair adheres to it due to a shared static electric charge. Similarly, when the voltage increases, the electrodes draw closer, causing the oil inside the bag to be displaced to one side, thereby shortening its overall length.
The paired actuators, when connected to a skeletal framework, simulate identical paired muscle movements found in living organisms; as one muscle contracts, its corresponding counterpart relaxes. The researchers employ a PC code that interfaces with high-voltage amplifiers to control the contraction or relaxation of specific actuators.
Researchers compared the vital energy efficiency of their robotic leg featuring a novel mechanical design with that of a conventional robotic leg driven by an electric motor, highlighting the potential advantages of their innovative approach. While examining various concerns, they investigated how much energy is wastefully converted into heat. As evident on the infrared image, the motorized leg appears to expend significantly more energy when tasked with traversing an uneven terrain. The temperature inside the electro-hydraulic leg remains constant. Because the ostensibly pseudomuscular structure exhibits electrostatic properties. “When discussing sticky situations, Buchner draws an apt analogy: ‘Like when hair gets caught on a balloon, it can linger there for quite some time.'” Typically, electric motor-driven robots necessitate thermal management, prompting the need for additional heat sinks or fans to dissipate heat effectively into the surrounding air. Without explicitly stating so, our system would not need additional input from them, according to Fukushima’s observation.
The robotic leg’s ability to leap depends on its capacity to rapidly and explosively generate force to overcome its own weight. Researchers further validated that the robotic leg exhibits an exceptional level of adaptability, a crucial characteristic in soft robotics applications. Can the musculoskeletal system’s inherent elasticity enable flexible adaptation to the terrain in question? It’s no wonder that living creatures are vastly distinct. “When you refuse to adapt, even simple actions like walking on an uneven surface become significantly more challenging,” Katzschmann remarks. Consider simply stepping down from the pavement onto the street.
Unlike electrical motors that necessitate sensors to continuously monitor the robotic leg’s angular position, the artificial muscle adjusts its location through interaction with its environment. Two small buttons control the movement of this mechanical component: one for flexing the joint and another for extending it further. In Fukushima’s words: “Understanding how to adapt to the local topography is crucial.” As individuals land after jumping, they don’t need to anticipate beforehand whether they’ll require a 90-degree or 70-degree knee bend. Similarly, a robotic leg’s musculoskeletal system adapts seamlessly upon touchdown, adjusting its joint angle in response to the surface’s texture – firm or soft?
While the analysis focus on electrohydraulic actuators has been relatively recent, emerging only around six years ago. While the field of robotics has made significant strides with advancements in control systems and machine learning, a notable lag exists in the development of robotic hardware, equally crucial for overall progress. “This publication serves as a powerful testament to the enormous potential for transformative innovation that arises from the introduction of novel hardware concepts, such as the utilization of synthetic muscles,” Keplinger remarks. While Katzschmann suggests electro-hydraulic actuators may not be suitable for heavy equipment on construction sites, they offer distinct advantages over traditional electric motors. The importance of customization becomes glaringly apparent in applications akin to grippers, where the required actions must be tailored according to the specific object being grasped, such as a sphere, an oval, or a fruit like a tomato.
While Katzschmann acknowledges some limitations, his primary concern is that “our system remains relatively restricted in comparison with strolling robots powered by electrical motors.” Currently, the leg is tethered to a rod, causing it to jump in circular motions and limiting its ability to move freely; future research should aim to overcome these constraints, ultimately paving the way for the development of autonomous walking robots equipped with advanced artificial muscles. “He further explains that combining the robotic leg with existing quadruped or humanoid designs could potentially lead to the development of a battery-powered rescue robot.”
