Legged robots have gained popularity due to the use of electric motors, providing controlled movement for robotic limbs. However, when it comes to achieving the instantaneous power of biological muscles, electric motors fall short. To address this, researchers from the University of Cape Town have turned to an old-school alternative—pneumatics. Inspired by the rapid maneuvers of cheetahs, they have demonstrated that pneumatics can offer high force-to-weight ratios, built-in compliance, and cost-effective solutions, making them a viable option for powerful and agile legged robots.
The use of hydraulics in legged robots can be complicated, expensive, and messy if accidents occur. Furthermore, while hydraulics are easier to model and control, they lack forgiveness in real-world applications. Pneumatics, on the other hand, offer simplicity, relatively low cost, and compliance that hydraulics lack. The challenge lies in controlling pneumatics effectively due to the compressibility of air, which disrupts traditional control methods. Nevertheless, researchers have found that complex control may not be necessary to mimic cheetah-like locomotion.
Amir Patel, an associate professor at the University of Cape Town, who has conducted extensive research on cheetah biomechanics, believes that cheetahs do not rely on fine force control during rapid acceleration. Instead, they simply exert maximum force to achieve explosive motion. Patel’s team built a legged robot called Kemba to explore rapid acceleration and maneuverability using pneumatics. Kemba features high-torque electric motors for precise control at the hips, while pneumatic pistons controlled by binary valves provide explosive actuation at the knees.
Through extensive modeling of pneumatic dynamics, the researchers achieved explosive actuation with the pneumatics while obtaining finer control from the electric motors at the hips. Kemba, weighing 7 kilograms, was able to repeatedly jump to 0.5 meters with controlled landings and reach a maximum jump height of 1 meter. Although metrics like jump height and top speed are intriguing, Patel emphasizes the importance of focusing on the transient phase of locomotion—such as rapid acceleration and deceleration.
While Kemba is currently tethered, the researchers envision it becoming a platform for studying the biomechanics of animal locomotion. Upgrades such as a spine and tail are in consideration to enhance its degrees of freedom and enable more dynamic behaviors. While legged robots have a long way to go before matching the capabilities of real cheetahs, the pneumatic approach holds promise. Additionally, reducing the cost of legged robots could expedite their development, bringing us closer to a future with advanced legged robotic systems.
By Impact Lab