Meta is Planning to Develop an AI-Based Touch-Sensitive Robotic Gadgets with Artificial Intelligence

Dec, 2021 - By WMR

The Meta organization is developing a new type of artificial intelligence (AI) in robotics and exploring new fields of touch sensitivity.

Nowadays, humans are turning into robots and robots more into humans, recently renamed Meta is also joining this developing sector. The company is generally interested in robotics devices consisting of touch-sensitive technology.

Recently, Meta announced a new technology involving sensors to develop in the tactile sensing ecosystem. The sensation of touch is a fundamental element of encountering the surrounding environment. This is a sector in robotics that intends to learn and replicate human-level touch in the real world. Among certain progressions, robots will be extra effective in associating with the environment around people, making them much competent. The simulators, libraries, hardware, benchmarks, and datasets are the four columns of the touch-sensing ecosystem.

For this system, Meta presents DIGIT, a compact and high-resolution tactile sensor for in-hand manipulation. DIGIT is touch-sensing hardware. Meta AI declared the partnership with GelSight to commercially produce DIGIT, further making it convenient for researchers to assist them in taking out touch-sensing studies which offer speedy improvements.

Also, Meta's open-sourced TACTO is a simulator for high-resolution vision-based physical sensors which allows a more active experimentation program, giving assistance to ML researchers, in the absence of a hardware system. Moreover, TACTO can provide readings of high-resolution touch sensations at numbers of frames per second; this allows researchers to produce simulations of vision-based tactile sensors with various parts that can be installed on robots.

Along with DIGIT and TACTO, Meta AI is in partnership with researchers from Carnegie Mellon to produce ReSkin. The ReSkin is an artificial skin for robotic hands. ReSkin is a deformable elastomer including fixed shreds of magnets.

Meta’s simulators TACTO along with ReSkin and DIGIT with their datasets, it can make tactile knowledge available to researchers. This is taking AI behind its ability of see and sound. Nevertheless, despite Meta AI’s progress, there is a significant job still to be made in the tactical sensing sector.