The neuromorphic chip significantly reduces the energy requirements of the rolling robot
A team of researchers from the Center for Brain-Inspired Computing Research at Tsinghua University in Beijing, China, has developed a neuromorphic chip capable of roughly halving the energy consumption of a cat-like rolling robot and mouse, compared to a classic NVIDIA. chip designed for AI applications. In their article published in the journal Scientific Robotics, the group describes the design concepts they used to build the chip and how well it performed in testing.
As AI applications mature, they become more complex and require more power to operate, which can be a problem for autonomous robots working in the field. Most of these systems are based on the use of neural networks. In this new effort, the researchers believed that building a chip with similar capabilities but based on neuromorphic computing technology would use much less power. They built a neuromorphic chip called TianjicX and put it inside a small rolling robot called Tianjicat.
The chip was imbued with a spatio-temporal elasticity that allowed adaptive allocation of its resources and also the scheduling of multiple tasks (the robot had to be able to track the mouse and pursue it, while processing and responding to information about obstacles). It also had a high-level module that bridged the gap between the given requirements and the physical architecture of the robot.
Tianjicat was then programmed to track a target and avoid hitting obstacles using data from onboard sensors. The target in this case was a remote-controlled rolling toy with a cartoon mouse attached to the top. The researchers called it a cat-and-mouse challenge.
The researchers found that the mechanical cat was quite capable of chasing mice while avoiding obstacles. He was also able to catch mice. Tianjicat used just over 50% less energy than an identical robot based on an NVIDIA chip. And they also found that their neuromorphic chip-based robot had significantly reduced latency – 79 times less than the NVIDIA-based system, allowing the robot to make decisions much faster.
A neuromorphic computing architecture capable of running certain deep neural networks more efficiently
Songchen Ma et al, Spatiotemporal Elastic Neuromorphic Computer Chip for Intelligent Multitasking Robots, Scientific robotics (2022). DOI: 10.1126/scirobotics.abk2948
© 2022 Science X Network
Quote: Neuromorphic chip dramatically reduces rolling robot power requirements (Jun 16, 2022) Retrieved June 16, 2022 from https://techxplore.com/news/2022-06-neuromorphic-chip-power-requirements-robot.html
This document is subject to copyright. Except for fair use for purposes of private study or research, no part may be reproduced without written permission. The content is provided for information only.