GM: AI in memory for autonomous driving

With Untether AI, the manufacturer General Motors is developing an improved AI for sensor evaluation.

The company Untether AI develops AI inference calculations that run in memory. You now want to work with the manufacturer for this General Motors work together. The project is supported by the Ontario government and aims to take the perception systems of autonomous driving to the next level.

According to Untether AI, more than 90% of the AI’s power consumption comes from the movement of data. Untether AI moves the compute element to where the data resides—in memory. This will make the power consumption for data transmission is reduced by a factor of six. This is the innovation by which the computational density can be provided.

The heart of the unique at-memory computing architecture is the memory bank 385 KB SRAM with a 2D array of 512 processing elements. With 511 arrays per chip, each device offers 200MB of memory, enough to run a few networks on a single chip.

The agreement between Untether AI and GM for the realization of such a system for autonomous driving is signed by ontario supported with one million US dollars. The declared goal of the project is to increase the performance of autonomous vehicles and reduce the power consumption of the systems, which also increases battery life. Overall, one also wants to reduce the costs for autonomous driving.

Go to Source