Feel the Future: Robotiq Gives Physical AI a Sense of Touch
Robotiq has officially bridged the gap between vision and touch with the launch of its TSF-85 Tactile Sensor Fingertips. Designed for the industry-standard 2F-85 Adaptive Gripper, these fingertips provide Physical AI systems with the high-frequency sensory feedback necessary to understand and interact with the physical world at scale.
While vision-based systems allow robots to see, true manipulation requires contact awareness. Without the ability to feel force and friction, robots struggle with complex tasks. The TSF-85 addresses this by enabling robots to perceive contact geometry, allowing for more reliable interaction across diverse environments and objects.
"Physical AI demands reliable interaction with the real world," says Vincent Duchaine, CTO of Artificial Intelligence at Robotiq. By integrating high-frequency tactile sensing into an adaptive mechanical framework, Robotiq offers a way for robots to generalize tasks without the prohibitive cost and fragility of anthropomorphic hands.
VIDEO: Robotiq 2F-85
The 2F-85 base is built on a patented design that goes beyond traditional parallel grippers. Its ability to perform both pinch and encompassing grips allows it to conform to varying object shapes. This inherent mechanical adaptability reduces the complexity of grasp planning and lessens the system's total reliance on perfect visual data.
The new tactile fingertips introduce a sophisticated sensing layer, featuring a 4×7 static taxel grid to monitor force distribution. For high-stakes manipulation, the sensors offer micro-slip detection at 1000 Hz, ensuring that the robot can adjust its grip in real-time to prevent dropped objects or crushed materials.
Beyond force and slip, the fingertips include an integrated IMU for proprioceptive sensing. This combination allows AI models to gather rich datasets regarding contact awareness. These capabilities are essential for training reinforcement learning (RL) and vision-language-action (VLA) models that require precise sensory input.
Unlike custom-built tactile hands that are often too fragile for industrial use, Robotiq's solution is engineered for long-term deployment. With thousands of grippers already operating globally, the company emphasizes uptime and a low total cost of ownership compared to experimental, DIY tactile systems.
Integration is streamlined for rapid scaling. The fingertips utilize native RS-485 communication and a USB conversion board, making them compatible with various robot brands and research platforms. The hardware is designed to preserve the gripper's full stroke and reach while protecting internal cabling from the rigors of daily operation.
Robotiq's approach provides a practical bridge from the lab to the production line. By standardizing hardware and tactile data across research campuses, teams can move faster from validation to large-scale deployment. This consistency is a major advantage for organizations building massive Physical AI datasets.
The TSF-85 tactile fingertips represent a significant step forward in the evolution of General Purpose Robots. By giving machines a sense of touch that matches their visual capabilities, Robotiq is enabling the next generation of AI to learn faster, operate more robustly, and move beyond the laboratory.
Want more information? Click below.
Rate this article
View our terms of use and privacy policy