2025年11月27日，福莱新材在上海举行第三代电子皮肤新品发布暨投资者交流会，推出集“柔性材料、芯片、算法与大模型”于一体的新一代触觉传感系统。此次产品升级以“芯感一体”为核心方向，推动机器人触觉从采集原始信号进一步走向本地处理和智能识别。
触觉系统从“摸到”迈向“摸懂”
传统触觉传感方案主要负责采集压力或电阻变化，后续计算通常由外部控制器完成。第三代产品将信号处理能力更靠近传感端，使触觉数据能够在机器人指尖或末端快速处理，减少数据传输和上位系统决策带来的延迟。
公开报道将这一触觉智能处理单元称为TPU，即Tactile Processing Unit。这里的TPU面向触觉数据处理，通过边缘计算支持快速反馈，使机器人不仅知道“是否接触”，还能够进一步判断接触物体的状态和特性。
三项核心方向构建触觉智能
芯感一体：将信号处理芯片与柔性传感器更紧密地集成，让算力下沉到传感端，缩短信号从采集到响应的链路。在抓取物体发生滑移等瞬时情况下，本地快速处理有助于系统更及时地调整动作。
算法赋能：在触觉数据处理中加入纹理识别、软硬感知和滑移检测等算法，将原始电信号转换为机器人更容易使用的物理属性，为物体识别、抓取策略和安全交互提供信息。
全域触觉：覆盖范围从指尖逐步延伸至指腹、手掌、手背及机器人其他关键部位，在保持柔性贴合和多区域感知的同时，构建更完整的触觉安全边界。
材料、器件与算法协同升级
第三代产品不是单一传感器参数的升级，而是材料配方、柔性结构、采集芯片、算法与数据应用的系统整合。柔性材料负责适应曲面和形变，传感单元采集接触信息，芯片承担信号处理，算法则将数据转化为可用于控制和认知的结果。
这种一体化方向有助于降低机器人厂商的集成复杂度，但不同灵巧手结构、接口和任务仍需要针对性适配。产品后续还需经过客户验证、应用测试和规模化交付检验。
面向具身智能与多场景应用
除灵巧手和人形机器人外，相关能力还可用于工业检测、智能穿戴、医疗健康及人机交互等领域。公司表示，将继续围绕柔性材料、传感器、芯片与算法构建触觉解决方案，推动触觉从单点感知向系统智能演进。
本文为官网根据公司发布和公开媒体报道整理，具体技术指标与商业进展以公司正式资料为准。
福莱新材官方发布 ｜ 中国证券报·中证网报道
On November 27, 2025, Fulai introduced a new-generation tactile sensing system integrating flexible materials, chips, algorithms and large models at a product launch and investor event in Shanghai. Centered on chip-sensor integration, the upgrade advances robotic touch from raw-signal acquisition toward local processing and intelligent recognition.
From detecting touch to understanding touch
Conventional tactile systems mainly capture pressure or resistance changes, while subsequent computation is performed by an external controller. The third-generation product moves signal processing closer to the sensor, enabling rapid processing at the fingertip or end effector and reducing transmission and host-decision latency.
Public reports refer to the tactile processing unit as a TPU. In this context TPU means Tactile Processing Unit: an edge-computing component for tactile data that supports rapid feedback so the robot can determine not only whether contact occurred but also properties and state of the contacted object.
Three core directions for tactile intelligence
Chip-sensor integration: tighter integration of signal-processing chips and flexible sensors moves computation to the sensing edge and shortens the path from acquisition to response. Local processing can support faster correction when an object begins to slip.
Algorithm enablement: texture recognition, compliance perception and slip detection convert electrical signals into physical properties that a robot can use for object recognition, grasp planning and safe interaction.
Full-area touch: coverage expands from fingertips to finger pads, palms, the back of the hand and other critical robot areas, creating a more complete tactile safety boundary while maintaining conformability and multi-region sensing.
Coordinated advances in materials, devices and algorithms
The third-generation product is a system-level integration of material formulation, flexible structure, acquisition chips, algorithms and data applications rather than a single sensor-parameter update. Flexible materials conform to curved surfaces, sensing cells acquire contact, chips process signals and algorithms turn the data into control and cognition inputs.
This integrated approach can reduce implementation complexity for robot manufacturers, although every dexterous-hand structure, interface and task still requires adaptation. The product must continue through customer validation, application testing and scaled-delivery verification.
Embodied intelligence and broader applications
Beyond dexterous hands and humanoid robots, the capabilities may support industrial inspection, intelligent wearables, healthcare and human-machine interaction. Fulai states that it will continue building tactile solutions around flexible materials, sensors, chips and algorithms.
This article was compiled from company releases and public media reports. Official product materials govern technical specifications and commercial progress.
