Researchers from Tsinghua University have developed an autonomous self-driving bicycle that can even maneuver itself around and over obstacles. How? It’s powered by a Tianjic chip, which integrates two approaches to provide a hybrid, synergistic platform. In other words, it uses a multi-core architecture, reconfigurable building blocks and a streamlined dataflow with hybrid coding schemes that not only accommodates computer-science-based machine-learning algorithms, but also easily implement brain-inspired circuits and several coding schemes.
If you’re wondering how this differs from typical AI devices, this bicycle is just executing software that had been trained to handle specific tasks rather than learning in real-time. This includes recognizing spoken words and avoiding obstacles, which is essential for this bicycle, which runs on battery power. The goal is to seamlessly blend the training process and the in-the-moment execution, so that it could learn near instantly after just a few moments of experience.
No products found.
Using just one chip, we demonstrate the simultaneous processing of versatile algorithms and models in an unmanned bicycle system, realizing real-time object detection, tracking, voice control, obstacle avoidance and balance control. Our study is expected to stimulate [artificial general intelligence] development by paving the way to more generalized hardware platforms,” said the team of scientists.
