MIT News January 21, 2021
In complex situations robots often do not move quickly because perceiving stimuli and calculating a response takes a lot of computation which limits reaction time. A team of researchers in the US (MIT, Harvard University) used robomorphic computing to bridge the mismatch between a robot’s “mind” and body. Their system creates a customized hardware design to best serve a particular robot’s computing needs. The user inputs the parameters of a robot, the system translates these physical properties into mathematical matrices. These matrices contain many zero values that roughly correspond to movements that are impossible given a robot’s particular anatomy. The system then designs a hardware architecture specialized to run calculations only on the non-zero values in the matrices. The resulting chip design is therefore tailored to maximize efficiency for the robot’s computing needs. In tests the hardware architecture designed using this method for a particular application outperformed off-the-shelf CPU and GPU units. The researchers see widespread potential for robomorphic computing. The research will be presented at a meeting in April…read more.