Most people think big breakthroughs win robot races. This time, a tiny change did it. In a humanoid robot race across parts of the Gobi Desert, two near‑identical Star1 robots ran side by side. The only difference was footwear. The one with sneakers won. This humanoid robot race is a good reminder: in real‑world tasks, small design choices often make the biggest difference.
Humanoid robot race: what the Star1 test shows
In late 2024, Chinese start‑up Robot Era filmed two Star1 robots running on roads, grass, and rocky paths in the Gobi. One ran barefoot, one wore shoes. The sneaker‑equipped robot caught up and passed the other, reaching about 3.6 m/s for more than half an hour, according to media reports and the company’s own notes. The short video was a promo, not a lab trial, but it highlights a simple idea: contact with the ground is a major limiter for legged machines.
Why this matters for real use
Most future tasks for humanoids involve walking on mixed surfaces: factory floors, sidewalks, ramps, and steps. Grip, cushioning, and foot shape affect stability, energy use, and wear. The run suggests that tuned soles, materials, or even modular “shoes” could improve safety and speed without changing the whole robot.
Context, not a world record claim
Robot makers often compare speeds, but conditions differ. A fun run on uneven ground is not the same as a lab sprint on a flat track. Treat the Gobi video as a field demo of robustness, not a final scorecard.
Robot Era Star1 speed and specs, in simple terms
Star1 is a full‑size biped, about 1.71 m tall and 65 kg. The joints use high‑torque motors (up to around 400 N·m) and fast control. The company says the robot has an on‑board AI computer rated at roughly 275 TOPS (trillion operations per second). In plain words: strong joints and fast perception/planning help the robot keep balance while it changes from walking to running and back.
What we saw in the video
The robots keep a straight torso and swing their legs forward in a stiff, almost toy‑soldier style. It looks odd, but the posture may help on rough ground by keeping the mass centered while the feet search for grip. One segment even shows the sneakered robot jogging beside humans.
What the numbers do not tell
Top speed alone says little about usefulness. Range per charge, fall recovery, safe hand use, and simple service matter more. Also key is how the robot learns new moves and adapts to new floors or weather.
Why shoes mattered for the robot’s feet
A humanoid foot is a hard problem. Designers trade off traction, compliance (the “give” in the sole), and sensing. On soft or uneven surfaces, a slightly softer sole can spread pressure and maintain contact. On hard road, a stiffer setup can reduce wobble. The shoe likely added grip and gentle damping, reducing slips and impacts. That small change seems to have saved energy and boosted stability at speed.
Design ideas that follow
- Removable soles: swap tread patterns for sand, grass, tile, or ice.
- Embedded sensors: pressure maps in the sole to detect slip early.
- Passive compliance layers: thin foams or elastomers to smooth shocks.
- Toe and heel geometry: tuned roll‑over helps smoother push‑off.
Limits and what to watch next in humanoid robots
Evidence quality here is low: a company video and tech press coverage, not a peer‑reviewed study. Still, the demo fits with what we know about legged robots: ground contact is everything. Watch for repeatable tests, fall‑recovery demos, and long duty cycles under load. Also look for safer remote‑control methods and better human‑robot teaming. For example, new interfaces let operators control a humanoid robot with simple muscle twitches, which could speed training and raise safety.
Limitations & quality of evidence
- Evidence type: product demo video; secondary news reports.
- No independent timing or distance measures were shared.
- The “with shoes” factor could include other small set‑up changes we cannot see.
Futura Sciences – Dans cette course improbable de robots humanoïdes, un détail a déterminé le vainqueur ! – 2024
A French tech outlet reported that the sneaker‑equipped Star1 caught and passed the barefoot unit on mixed terrain in the Gobi. Evidence type: company video coverage.
New Atlas – Watch: Sneaker‑wearing humanoid beats barefoot bot on Gobi fun run – 2024
New Atlas summarized the test and noted the 3.6 m/s top speed during a 34‑minute run, plus basic Star1 specs. Evidence type: company video coverage.
Robot Era – STAR1 Gobi run video and product brief – 2024
The maker’s site hosts a video tile titled “STAR1 Gobi Desert run” and a spec sheet stating 400 N·m joint torque and on‑board AI up to 275 TOPS. Evidence type: manufacturer materials.
Gadgets 360 – STAR1 Robot breaks speed record with sneakers, reaches 8 mph – 2024
A news brief claimed the sneaker‑wearing Star1 reached about 8 mph (3.6 m/s) and ran for 34 minutes in the Gobi. Evidence type: secondary media.
0 Comments