Sabrina Neuman, a graduate of the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), and her colleagues have found a way to address the performance gap between a robot's computational power and its joint actuator response times. The method, called robomorphic computing, uses a robot's physical layout and intended applications to generate a customized computer chip that minimizes the robot's response time.
Neuman and her co-authors describe the approach in "Robomorphic Computing: A Design Methodology for Domain-Specific Accelerators Parameterized by Robot Morphology," which will be presented in April at the virtual ASPLOS 2021, the 26th International Conference on Architectural Support for Programming Languages and Operating Systems.
The robomorphic computing system creates a customized hardware design tailored to maximize efficiency for a robot's computing needs.
Hardware architecture designed using this method for a particular application outperformed off-the-shelf CPU and GPU units. While Neuman's team didn't fabricate a specialized chip from scratch, they programmed a customizable field-programmable gate array chip according to their system's suggestions. Despite operating at a slower clock rate, that chip performed eight times faster than the CPU and 86 times faster than the GPU.
From MIT News
View Full Article
No entries found