A group of researchers at the Massachusetts Institute of Technology have developed a novel system to design hardware architecture on the go to improve a robot’s performance in its respective domain. The robomorphic computing system gauges the physical layout of a robot and also determines the tasks to be performed to suggest an optimized architecture for the hardware. Robomorphic computing gauges the robot’s physical layout and tasks to be performed to suggest an optimized hardware architecture.
The main component of a robot are motors coupled with the latest technologies of Artificial Intelligence, IoT, etc. This facilitates the faster movement of robots in addition to automation. Yet, the computations to initiate action continues to be a drag thereby causing a delay in its reaction time, especially in complex situations like interacting with humans where errors are least preferred. The researchers have thus come up with robomorphic computing to accelerate this reaction time.
WORKING:
A robot’s operation can be summarized in three simple steps:
1. Data gathering using cameras and sensors
2. Understanding the environment using mapping and localization techniques
3. Plotting the course of action
Embedded vision and SLAM technology make data gathering as well as localization easier. Yet, all these steps are extremely time-consuming, especially when the calculations are done on a CPU. Earlier the researchers investigated the software side to speed-up robots, following which they decided to approach the problem from the hardware side. Hardware acceleration is defined as the ‘use of a specialized hardware unit to do certain computing tasks more efficiently’. While Graphic Processing Units or GPUs have been used for such tasks, the application is limited as the use case varies from robot to robot.
Robomorphic computing has been developed to put together a customized hardware unit for individual robots. This method of edge computing takes into consideration the physical parameters of the robot and the tasks it needs to perform. These are then translated into mathematical matrices which results in the design of customized hardware architecture, intended to increase the efficiency of robots, with a unique chip design. The researchers also claim that a chip developed using robomorphic computing performed eighty-six times faster than a CPU and eight times faster than a GPU.
ADVANTAGES:
· The possibilities in this field of computing are immense now given that a safe, customized planning chip, unique to each robot is greatly efficient as well as cost-effective.
· Robomorphic computing can significantly bring down the designing cost of customized chips.
· Robomorphic computing ensures smooth, safe, efficient, and faster functioning of robots among humans.
· Better scope for the use of customized robots to enter zones potentially inaccessible to humans to facilitate research purposes, etc.
The researchers at MIT plan to automate the entire process of robomorphic computing. This will facilitate users to simply input parameters to manufacture the hardware designs. Today, highly automated robots have found their way into almost every sector with different applications. Robomorphic computing is truly a game-changer in the field of robotics, empowering users to manufacture customized robots, affordably.
Follow and connect with us on Facebook, Linkedin & Twitter
