In a groundbreaking yet humbling debut, humanoid robots struggled to keep up with human runners in Beijing’s half-marathon, highlighting design flaws and limitations in their AI technology.
A group of 21 ‘humanoid robots’ participated in a half-marathon race in Beijing on Saturday, with only six completing the course. The event, which saw human athletes running on separate tracks, was organized by several branches of Beijing’s municipal government and marked the first time humans and bipedal robots have run together.
Humanoid robots are designed to resemble humans in appearance and movement.
The first humanoid robot, Wabot-1, was developed in Japan in 1972.
Since then, advancements have led to the creation of more sophisticated models like Honda's ASIMO and Boston Dynamics' Atlas.
Humanoid robots are used for search and rescue missions, healthcare, and space exploration.
They possess advanced features such as balance control, speech recognition, and dexterous manipulation.
The fastest robot, Tiangong Ultra, finished the race in 2 hours and 40 minutes after its battery was changed three times and it fell down once. However, despite its impressive speed, the robot was unable to keep up with human runners, barely qualifying for a human participation award.
According to Alan Fern, a robotics professor at Oregon State University, researchers who build humanoid robots typically focus on completing tasks and responding effectively in diverse environments rather than running as fast as possible. The AI technology used in humanoids has not progressed much since 2021, when Fern’s team sent a bipedal robot to run a 5K race.
The event demonstrates the robustness of humanoid hardware, but also highlights the limitations of these robots. Almost every robot fell and faced overheating problems, prompting operators to switch them out for new replacements. The performance of the other humanoids was mostly disastrous, with some experiencing mechanical failures or struggling to maintain a steady pace.
Design Flaws Exposed
The half-marathon revealed several design flaws in humanoid robots, including inadequate cooling systems, fragile parts that can be easily damaged, and limitations in their ability to adapt to changing environments. The use of duct tape proved to be an effective problem-solving tool, with humans using it to repair broken robots and keep them running.
Design flaws can lead to product failures, safety issues, and financial losses.
Common design flaws include poor user experience, inadequate testing, and insufficient quality control.
Inadequate research and development can also result in design flaws.
According to a study, 60% of products fail due to design errors.
Identifying and addressing design flaws early on is crucial for product success.
Manufacturers must prioritize rigorous testing, customer feedback, and continuous improvement to minimize the risk of design flaws.
Despite these challenges, some robots were able to perform impressive tasks, such as dancing and backflips. However, these capabilities do not necessarily demonstrate the usefulness of humanoid robots in real-world situations. As Fern notes, ‘people are going to be bored of dancing and karate‘ and will instead want to see robots that can handle diverse real-world tasks without human instructions.
A New Direction for Humanoid Robots
The Beijing half-marathon marks a turning point for China’s humanoid robot industry. While the event highlighted the limitations of these robots, it also demonstrated their potential for innovation and improvement. As Fern predicts, ‘China shifting this year to focusing more on doing useful things‘ will be key to advancing the field.
The humanoid robot industry has experienced significant growth in recent years, driven by advancements in artificial intelligence, machine learning, and robotics.
According to a report by Grand View Research, the global humanoid robot market size was valued at USD 1.4 billion in 2020 and is expected to reach USD 6.2 billion by 2027, growing at a CAGR of 22.5%.
Humanoid robots are being used in various industries, including healthcare, manufacturing, and service sectors.
For now, however, the focus is on improving design flaws and increasing robustness. With over 12,000 human athletes competing in the half-marathon, it’s clear that humanoid robots have a long way to go before they can keep up with their human counterparts.
