Hubei Signs Strategic Agreement To Establish Robotics Industry's First Open-Architecture Tactile Cognitive Center
Leave a message
On May 5th, in the Wuhan East Lake New Technology Development Zone, the Hubei Humanoid Robot Innovation Center, Wuhan Huaweike Intelligent Technology Co., Ltd., and the Hubei Humanoid Robot Industry Alliance signed a strategic cooperation agreement to jointly establish the industry's first open-architecture, ultra-large-scale multimodal tactile data engine and cognitive training center. The center will systematically collect and train robotic tactile data, addressing the long-standing industry challenge of fragmented tactile data and enabling robots to "perceive the world."
Why Has Tactile Sensation Become the Focus of Humanoid Robot Development?
The industry is increasingly focusing on tactile sensation for humanoid robots. Recently, a user on X platform revealed what appears to be a prototype of Tesla's Optimus 3 humanoid robot. Previously, it demonstrated actions like "catching a tennis ball" that rely on real-time tactile feedback, leading netizens to speculate that this robot may incorporate electronic skin in multiple areas to build its tactile system.
The reason tactile systems for humanoid robots have garnered significant attention from global key enterprises lies partly in the diverse and complex nature of human tactile perception. Tactile sensation has always been a core component of the human sensory system, providing crucial input for interactive perception and automation systems in humanoid robots. It enables more accurate detection of an object's position and surface properties. On the other hand, after training humanoid robots to integrate multi-source sensor data like vision, touch, and force perception, deep learning models enable cross-modal associations and deep fusion of multi-modal data. This endows humanoid robots with human-like cognitive characteristics and facilitates cross-scenario transferability, representing a crucial direction for advancing their capabilities to the next stage.
From a market perspective, investing in tactile technology also represents a bet on future growth. According to Precedence Research data, as the core of humanoid robot perception, the global flexible tactile sensor market is projected to reach $31.06 billion by 2034.
What are the challenges in building a platform for tactile data acquisition and training?
The development of a tactile data acquisition and training platform represents a highly complex, interdisciplinary systems engineering endeavor. Its difficulty far exceeds that of building platforms for modalities like vision or speech. Establishing a tactile data perception system for humanoid robots is no simple task. It requires deeply integrating cutting-edge achievements from mechanical engineering, computer science, cognitive science, materials science, and other fields to form a closed-loop system spanning data collection, algorithm optimization, and hardware adaptation. Only then can this goal be truly achieved. This has drawn significant attention to Hubei's newly established tactile training center. The project constructs a "1+10+1000+N" tactile perception platform system (1 embodied tactile perception architecture / 10+ key scenarios / 1000+ common objects / N categories of real-world generalized applications). Through innovation across four dimensions, it aims to address multiple cutting-edge challenges in robotic tactile systems and advance the dexterous manipulation capabilities of humanoid robots.
This tactile training center integrates multimodal tactile data acquisition, adapted for over 10 key scenarios including healthcare, retail, industry, and laboratories. It has established tactile data collection for over 1,000 real-world objects, constructing a foundational tactile database for the physical world. This provides rich, authentic tactile samples for AI training across manufacturers, enabling robots to perform precise operations and environmental interactions. Equivalent to mapping a "tactile landscape," it supports the cross-industry and cross-scenario deployment of humanoid robots, while also laying the groundwork for the center's long-term, large-scale operations.
