A humanoid robot developed by PaXini Technology (Tianjin) Co is showcased for household application scenarios. CHINA DAILY

On sorting lines in smart factories, at precision workstations in automobile manufacturing plants and even at home tending to household chores, a batch of robots equipped with multidimensional tactile sensors can dexterously grasp, operate and adaptively perform tasks, accomplishing movements that were once only possible for humans.

These robots, capable of sensing pressure, texture and posture, possess "tactile fingertips", enabling artificial intelligence to transition from virtual algorithms to the real world. This is the industrial transformation brought about by embodied intelligence, and the core "sensory organs" enabling it all originate from Tianjin.

"Robots are important carriers of AI, facilitating its transition from the virtual world to reality," said Yin Jihui, secretary of the Hexi district Party committee. As a national advanced manufacturing research and development base, Tianjin is focusing on core technology research and industrial implementation of embodied intelligence robots, cultivating a group of globally competitive industry leaders. PaXini Technology (Tianjin) Co is one such representative example.

"The key to these logistics robots lies in our independently developed multidimensional tactile sensors," Lin Tong, director of PaXini' s data collection center, said. Currently, 80 percent of humanoid robots worldwide utilize PaXini's sensors. Tactile sensors, known as the cornerstone of the intelligent cognition and information interaction industry, have also been listed by China as one of the key critical chokepoint technologies alongside lithography machines and chips.

"Five or six years ago, similar tactile sensor products cost as much as 100,000 yuan ($14,770) each, severely restricting the development of the entire robot industry," Nie Xiangru, co-founder of PaXini, recalled. The high import costs made tactile perception a luxury in the robotics field, making it difficult to be commercialized on a large scale.

To break the monopoly, the PaXini team spent four to five years in R&D, successfully creating the world's first 6D hall array multidimensional tactile sensor, transforming core technology from a heavy reliance on imports to 100 percent domestic self-sufficiency.

Today, PaXini's basic tactile sensor models are priced at only 199 yuan, just 0.2 percent of the price of equivalent products in overseas market. This has made tactile perception accessible from laboratories to various industries, turning a luxury into a basic hardware accessible to everyone.

The key to reducing costs lies in significant optimizations in chips, structure and algorithms with each product generation, Nie said. "The advancement of software algorithms enables us to reduce our reliance on hardware. Coupled with the growth of the industry, when prices become cheaper, the order volume increases, and with more orders, the costs can be further reduced. Naturally, the prices spiral downward."

PaXini's competitiveness lies not only in the independent R&D of core components, but also in its underlying data.

Last year, PaXini's Super EID Factory, officially commenced operations in the Kongtian digital industrial park in Hexi district. It is currently the world's largest embodied intelligence data collection and model training base.

"For robots to be deployed effectively, they require sufficient scenario-based data support," Nie said. The factory utilizes critical data from various vertical work scenarios to enable robots to learn how to work, thus enabling them to adapt to practical applications more quickly.

As the world's largest EID collection and model training base, this factory's daily collection capacity is equivalent to 1.02 to 1.52 times that of an entire worker's career. The factory encompasses thousands of tasks across full-chain embodied intelligence application environments, including automotive manufacturing, households, offices, catering, supermarkets, healthcare and medical services, aiming to simulate the complex scenarios robots may encounter in the real world.