SCHEDULE BUILDER
Welcome to the Drive World with ESC 2019 Presentation Store. Here you can view and download conference presentations before, during, and after the event. If you’re looking for a presentation from a specific session that you’re unable to find here, note that it’s likely because the presenter has not provided permission for external use or has not yet shared their presentation with us. Please check back after the event for a more complete catalogue of available presentations.
Endpoint AI for the IoT – How Much Computing Power Do You Need?
Charles Lord (President, Blue Ridge Adv Design)
Location: Room 203
Date: Tuesday, August 27
Time: 9:00am - 11:00am
Track: Tutorials & Hands-on Demos, ESC - Advanced Technologies
Format: Tutorial/Demo
Vault Recording: TBD
With the explosion of the Internet of Things (IoT), more and more data is being generated and must be acted upon. Even with high-throughput channels such as 5G, more and more processing power will be needed at IoT endpoints to reduce traffic. One solution that is gaining popularity is the implementation of artificial intelligence (AI), particularly neural networks (NN) in order to recognize and prioritize data at the endpoint. But at what cost of processor resources? We will take a look of the capabilities of various ARM processor technologies from the simple Cortex-M to multicore specialized processors such as the Renesas A2 and the NVIDIA Jetson with some real-life examples. We will step through the process of gathering data sets, generating learning models, and making those models fit the capabilities of the various sized processor technologies listed. We will then look at the performance capabilities of the above when coupled with the other processing needs of an IoT end node (for example, communications).