Innovative transistor saves circuits and delivers more performance

Breakthrough in edge computing: A new transistor enables more efficient and intelligent devices. Researchers are developing compact hardware that saves a lot of circuitry. This could boost the performance of smartphones, robots and autonomous vehicles.

Breakthrough: mini chip with mega performance

Edge computing refers to data processing directly at the point of origin, for example in smartphones or sensors, rather than in remote data centers. A team of researchers has now developed a new type of transistor that makes this technology much more efficient. The so-called “multi-gate van der Waals interface junction transistor (vdW-IJT)” is based on molybdenum disulfide (MoS2), a material with special electronic properties.

The transistor enables the implementation of fuzzy logic systems in compact devices. Fuzzy logic is a method that allows computers to deal with fuzzy or inaccurate information – similar to the human brain. Conventional computers work with binary logic, while fuzzy logic allows intermediate values ​​and can therefore make more complex decisions.

A particular advantage of the new transistor is its ability to directly generate certain mathematical functions (intrinsically Gaussian or n-shaped membership functions) that are important for fuzzy logic calculations. This eliminates the need for additional circuitry – traditional CMOS technology requires dozens of transistors to realize similar functions – thereby saving space and energy. In tests, the new hardware combined with artificial neural networks showed impressive results in tasks such as image recognition.

On the way

Like the researchers in Nature report, this technology could be used primarily in areas such as robotics and autonomous driving. It is important to carry out complex calculations quickly and energy-efficiently directly in the device. The researchers are now working on further developing the technology for mass production. If this succeeds, we could see significantly more powerful and energy-efficient intelligent devices in our everyday lives in the future.