Smart obstacle detection gets boost from multicore circuit

Inria’s Montbonnot lab (near Grenoble) recently developed an obstacle detection application capable of operating in dynamic environments—even going so far as to predict the trajectories of temporarily-hidden objects! The CEA, with the support of Inria, helped get the application—initially developed on a powerful workstation—ready for integration into a smart vehicle by embedding it into a multicore circuit by STMicroelectronics. The power-to-performance ratio improved tenfold and additional progress is within reach.
This type of ultra-compact, ultra-powerful circuit could potentially be used for other systems in fields like mobile medical imaging equipment and smartphones. However, Inria and the CEA are currently maintaining their focus on automotive applications. The research is being conducted under the Nanoelec research institute’s Bridging Technologies program.

Contact: diego.puschini@cea.fr

Facebook

Twitter