Skateboard Robot: From Sidewalk Surfing to ABB

A skateboard robot sounds like a stunt, but a rolling board is a hard balance problem. Four small wheels give almost no stability, and a rider stays upright by shifting weight all the time. This post follows the skateboard from California surf culture to the humanoid robots now learning to ride one. It looks at how researchers teach a machine to stay on a moving board, what a Japanese demonstration at CEATEC 2026 showed, and where ABB robots could fit. ABB has not built a skateboard robot, so that part describes possibilities and says so.

Skaters at the Pier 7 skate spot in San Francisco, 2022. Photo: Kylelovesyou / Wikimedia Commons (CC0)

How skateboards rolled out of the surf scene

Skateboarding began in California in the late 1950s, when surfers wanted something to ride on days without waves. The first boards were wooden planks or boxes with roller skate wheels bolted underneath. They were crude and hard to steer, and the clay wheels gripped poorly on rough pavement. By the early 1960s, companies were selling proper decks, and a 1960s Nash "Shark" board still shows the simple wooden shape of that era. The Shark deck in the picture is shaped like a small surfboard.

A 1960s style wooden Nash Shark skateboard deck shaped like a small surfboard. Photo: H. Michael Karshis / Wikimedia Commons (CC BY 2.0)

The big change came in the 1970s with polyurethane wheels. Frank Nasworthy introduced urethane wheels around 1972, and they gripped far better than clay and rolled smoothly over cracks and grit. Riders could turn harder, go faster and try tricks that were impossible before. In the same years, young skaters in Santa Monica and Venice, the Zephyr team called the Z-Boys, took the sport into empty swimming pools and steep banks, and board design followed them there. Wheel chemistry became a design topic of its own, and it still is.

A polyurethane skateboard wheel lit by the sun, the material that changed grip in the 1970s. Photo: Adam Rosenberg / Wikimedia Commons (CC BY-SA 2.0)

Skateboarding then moved into purpose built parks. Concrete bowls, rails and ledges gave riders a safe place to practise, and the sport spread to cities around the world. Every skater learns the same hard lesson: balance is a stream of tiny corrections, not a fixed pose. That is why the board attracts roboticists. Staying upright on a rolling platform needs fast sensing and smooth control, and that is exactly what a walking robot finds difficult. Parks gave the sport fixed, repeatable terrain that skaters and, later, test robots could return to.

A concrete skate bowl at Kenneli DIY in Tampere, Finland, 2025. Photo: Mikko J. Putkonen / Wikimedia Commons (CC BY 4.0)

Humanoids learn to ride, and where ABB fits

Robots have been on boards and wheels for some years. Caltech's LEONARDO, a bipedal robot, was shown riding a skateboard and walking a slackline. Unitree has shown its G1 humanoid on roller skates and ice skates, which tells you how quickly legged robots have improved. Most of these systems use machine learning trained in simulation, where a virtual robot can fall thousands of times without breaking anything. The trained controller is then copied to the real machine. Each new demo adds a trick, such as carving, jumping or recovering after a slip.

A humanoid robot standing on a show stage beside a robot dog at the Japan Mobility Show, 2025. Photo: RuinDig / Yuki Uchida / Wikimedia Commons (CC BY 4.0)

The newest example is a small humanoid that NEDO, Japan's New Energy and Industrial Technology Development Organization, showed at CEATEC 2026 in Chiba, held 13 to 16 October. According to Japanese coverage, the team captured a person riding a skateboard, converted the motion data to fit the robot's body, and trained it in a simulator. Preparing the data took about five days and training took about 24 hours. The result ran on the real robot without further training on it. The goal stated in the coverage is a robot that moves with flexible, human like adaptability near people.

An ABB IRB 1300 with Foundry Plus 2 protection, built to handle water and dust. Photo: ABB

ABB robots are not skateboarders, but the workflow will look familiar to anyone who programs them. Industrial teams build and test a cell in simulation first, for example in RobotStudio, before any metal moves. The NEDO team added noise and varied conditions during training so the robot could cope with real wheels and road surfaces. A factory cell faces the same gap between model and reality. As a possibility, ABB's compact IRB 1300 or the GoFa CRB 15000 cobot could hold boards for finishing or test rides. Programs like these are usually checked virtually, then moved to the real arm.

An ABB GoFa cobot loading parts into a machine at a small factory. Photo: ABB

Where skateboard robots go next

Humanoid robots are moving from stages and trade shows toward real tasks, yet riding a skateboard is mostly a research benchmark. It shows that a robot can handle a moving, unstable platform and recover from small errors. Those skills carry over to walking on uneven ground, stepping onto a moving walkway, or staying balanced while carrying a load. A home skateboarding robot is unlikely to be a product soon. The value lies in the control methods behind it, which can then be reused in other machines. That reuse is why a skateboard demo matters to engineers who never plan to build a skater.

Humanoid and quadruped robots on stage at the Japan Mobility Show, 2025. Photo: RuinDig / Yuki Uchida / Wikimedia Commons (CC BY 4.0)

The factory side is more practical. Skateboards are made from laminated maple veneers that are pressed, cut, sanded, printed and fitted with trucks and wheels. Much of that work is repetitive handling and finishing, the kind of job an arm such as ABB's IRB 6700 or a dual arm YuMi could take on, though no source describes ABB doing this today. Any such cell would need safety checks, which is what SafeMove is made for. Our guide to ABB robot use cases shows how similar jobs are set up. A cobot might also let a small workshop finish custom decks without a large line, but that is only an idea.

Early ASEA and ABB industrial robots on display, including the IRB 6 and IRB 2000. Photo: Dependability / Wikimedia Commons (CC BY-SA 4.0)

Open problems remain. Humanoids are still expensive, they fall more often outside the lab than in it, and their batteries run for limited periods. Safety is the largest issue, since a falling robot can hurt people nearby. Skills are another gap, because few engineers can build both the simulation and the hardware. Skateboarding itself became an Olympic sport at the Tokyo Games, and a robot that rides a board may one day help coach beginners, but that remains a research goal and not something you can buy. For now, the realistic gains are in control research, not in the toy aisle.

A skater leaps off a board in a kickflip. Photo: Almonroth / Wikimedia Commons (CC BY-SA 3.0)

Innovation

AI solutions for effortless ABB robot control.

Automation

Robotics

ceojohntran@chatwithrobot.net

+84905311611

© 2025. All rights reserved.

qtran1215@gmail.com