Read more from Author Rachel Maga here: https://globelivemedia.com/author/rachel-maga/
MIT researchers are now addressing the big gap between how fast robots can process information (still slow) and how fast they can move (very fast thanks to advances in modern hardware). I’m trying to do that, and I’m using what’s called “robomorphic computing.”
This method was devised by Dr. Sabrina Neuman, a graduate of the MIT Computer Science and Artificial Intelligence (CSAIL), and provides hardware acceleration as a means of speeding up response times. It’s about using a customized computer chip that can.
Custom-made chips that are customized for a particular purpose are not new. But more local computing on devices with more modest power and processing power constraints than companies and technicians moving data back and forth between large data centers and devices over network connections. Custom chips have become more common as we have come to demand that we do.
With this method of robot morphic computing, a super-specialized chip designed according to the physical layout and application of the robot will be manufactured. If the robot recognizes the surrounding environment, positions and understands itself in it, considers the planned behavior based on it, and complements the software algorithm with hardware acceleration, the efficiency of the final stage can be achieved. Researchers can design processing chips that will significantly improve.
A typical example of hardware acceleration that many people encounter on a daily basis is the GPU (Graphics Processing Unit). The GPU is basically a processor specially designed for image processing such as display rendering and video playback. Today, GPUs are widely used because almost every computer runs an image-intensive application. But lately, custom chips with a variety of features have become more common, thanks to the evolution of more customizable and efficient small lot chip manufacturing technology.
MIT News explains how Dr. Neumann’s system works, especially when optimizing the design of robot control hardware chips:
The system creates a customized hardware design that best suits the computing needs of a particular robot. The user enters the robot’s parameters, such as the layout of the robot’s limbs and how various joints move. Dr. Neumann’s system transforms these physical properties into mathematical arrays. These sequences are “sparse”, meaning that they contain many zero values that roughly correspond to movements that are not possible with a particular anatomy of the robot. (Similarly, your arms are restricted in movement because you can only bend at specific joints. It’s not an infinitely flexible spaghetti noodle).
The system designs a hardware architecture that specializes in computing only non-zero values in an array. The resulting chip design is therefore customized to maximize efficiency to meet the computing needs of the robot. This customization has been successful in testing.
Dr. Neumann’s team used an FPGA (Field-Programmable Gate Array) in the test. It’s like something between a completely custom chip and an off-the-shelf CPU, with significantly better performance than the latter. In other words, if you actually make a custom chip from scratch, you can expect a much greater performance improvement.
The fact that robots react faster to the environment does more than just increase the speed and efficiency of production (although it does, of course). It also means that robots can work more safely in situations where people are working right next to or working with them. This is a major barrier to the wider use of robotics in our daily lives. In short, Dr. Neumann’s research could help open the door to a sci-fi future where humans and robots live in harmony.
Rachel Maga is a technology journalist currently working at Globe Live Media agency. She has been in the Technology Journalism field for over 5 years now. Her life’s biggest milestone is the inside tour of Tesla Industries, which was gifted to her by the legend Elon Musk himself.