The Tiangong robot climbs steps.

The Tiangong robot has recently achieved a global first by successfully climbing 134 steps outdoors after being fully upgraded.

The Tiangong robot is well-known for its superhuman running capabilities and powerful generalized mobility in complex terrains. Recently, the National New Energy Vehicle Technology Innovation Center has continued to enhance its embodied cerebellum capabilities, achieving vision-based perception and walking.

Tiangong can now cross multiple levels of stairs and 35-centimeter high steps without collisions, stepping on edges, or missing steps. It can also reach a running speed of up to 12 kilometers per hour and can run on sand, snow, and slopes. Meanwhile, Tiangong has stronger anti-interference capabilities and can maintain balance even under significant external forces.

The improvement in mobility in complex terrains will enable humanoid robots to move out of the laboratory, perform tasks in real environments, and even operate in extreme environments such as mountains, snow, and ruins. This advancement lays a solid technological foundation for the large-scale application of embodied intelligent robots.

Tiangong robot

The Tiangong robot achieves "perception and walking" through the embodied "brain" and embodied "cerebellum" contained in the multi-functional embodied intelligent entity "Kaiwu". It can instantly "see clearly" the terrain ahead during movement, immediately generate accurate environmental analysis, dynamically adjust full-body movements and gait, and adeptly respond to randomly changing stair conditions.

With the upgrade in full-body control capabilities brought about by the embodied cerebellum, the Tiangong robot's mobility in complex environments has once again significantly improved. For the first time, it fully leverages the advantages of bipedal structure for humanoid robots in various terrains, establishing a new benchmark in the industry for complex environment mobility while achieving a closed-loop technology system in all-terrain scenarios.