Making better robots depends on better data capture, Chinese firm 51World says
The race to develop intelligent humanoid robots faces a major roadblock in a severe shortage of high-quality training data, but Beijing-based tech company 51World believes it has the tools to break the bottleneck. Best known for its digital twin and simulation technology, 51 World on Tuesday unveiled a new suite of data-collection devices and platforms designed to help train embodied AI systems –…
Chinese startup 51World has unveiled a new suite of data-collection tools to tackle the shortage of high-quality training data for intelligent humanoid robots. The company, known for its digital twin and simulation technology, aims to address one of the biggest hurdles in the embodied AI industry. Li Yi, 51World's founder and CEO, highlighted the importance of high-precision data in developing robots that can operate stably and reliably.
The company's hardware suite includes a headset, AperEgo, which captures a panoramic view of surroundings and synchronizes visual, motion, and audio inputs with high accuracy. AperOS platform processes the data, filtering out invalid information to reduce storage and computing costs. Li Yi believes that integrating hardware and software can push precision up to 99 percent, leading to greater accuracy and efficiency.
China's humanoid robot market is expected to reach 15 billion yuan (US$2.2 billion) by 2026, with a further 60 percent growth in 2027. 51World plans to power around 1,000 embodied AI-equipped robots with its platform by the end of the year. The Chinese government has been supporting the deployment of humanoid robots and embodied AI, unveiling a plan to boost high-quality AI training data supply.
51World's decade of experience in synthetic data and simulation platforms for autonomous driving helped pave the way for its expansion into humanoid robotics. Li Yi sees a parallel between the data collection approach used in autonomous driving and the need for humanoid robotics to follow suit for scalability. He emphasized that improving data costs, collection speeds, and generation efficiency is crucial to achieving artificial general intelligence.
The company is also looking to tap into the growing demand for data in the United States, where humanoid robotics and embodied AI are also gaining traction.
Written by urgent.news from SCMP Business's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
