Unmanned bicycle driven by AI

A research team at Tsinghua University in China developed an AI chip called Tianjic, which aimed at artificial general intelligence (AGI), or artificial intelligence, and mounted it on a bicycle. By doing this, it is said to have demonstrated real-time object detection and tracking, voice recognition, obstacle avoidance, and balance control. An unmanned bicycle that can be controlled by voice was born.

This bike runs sideways, avoiding undulations or obstacles. Humans are riding bicycles, but it is too complicated and difficult for computers. But thanks to AI, this is made possible through machine learning.

Tenjik is made in the form of an integrated core FCore, which combines artificial neural networks, neural network synapses, and blocks that act as cell bodies. Using a 28nm manufacturing process, the chip consists of 156 integrated cores and contains 40,000 neurons and 10 million synapses in an area of 3.8mm2.

The bicycle was equipped with an inertial measuring device, a camera, a steering wheel adjustment motor, a driving motor, an accelerator motor, and a rechargeable battery, and was controlled with a tanjik chip to enable decision-making by integrating various information. Chip itself is said to have been able to easily drive a bicycle even in practice after undergoing several training in advance.

The Tsinghua University research team said that although other road conditions, noise, weather factors, and multi-language support are already available in the current state, they have already developed a chip for driving a bicycle and are working on a next-generation chip research. Related information can be found here .