Chinese Engineers Created Wearable Robot Assistant For Heavy-Duty Tasks

Credit: SUS Tech

Chinese engineers from the Southern University of Science and Technology in Shenzhen have developed an unusual “wearable” robot that, when paired with a human, resembles a mythical centaur. It’s a platform for carrying heavy loads on off-road surfaces, with the robot’s legs supporting some of the load. The platform is controlled by the human but adjusts its stride to the user’s walking pace and direction.

The system is attached to the user’s back via an elastic interface. Unlike traditional exoskeletons, which rely on the user’s own legs for support, the Centaur robot acts as an independent pair of limbs. It supports vertical loads and assists with forward propulsion, while the user controls balance and direction.

The key solution is an elastic link with nonlinear stiffness. At low loads, the connection remains rigid, ensuring precise coordination. As the load increases, the system becomes more flexible: the robot absorbs shocks and absorbs most of the force. This approach allows for task separation: the user handles navigation, while the machine handles the mechanical work.

Volunteers were tested with a 20 kg load on their backs. The results, published in The International Journal of Robotics Research, are compelling:

• The robot supported 52.2% of the total load;
• The human’s metabolic expenditure decreased by 35.2%;
• The pressure on the feet was reduced by approximately half;

The gait remained stable, and the stride was predictable even on uneven surfaces.

The researchers noted that the system adapts to different walking speeds and directions. Predictive control and trajectory planning algorithms allow the robot to maintain balance and not interfere with human movements.

The developers see great potential for the Centaur in applications that require long-term heavy-duty transport: military logistics, emergency rescue operations, industrial work, and transportation over difficult terrain. This robotic loader doesn’t require complex “brains,” allowing it to be used effectively right now.