How to Choose an Edge AI Device

Choosing a device to perform edge AI tasks can be daunting due to the number of available options. We provide several factors to consider when choosing such a device, including use case, interface, power, form factor, environment, and code portability. We also examine the tradeoffs between buying off-the-shelf components or solutions versus building them yourself.

Edge AI devices generally fall into one of five categories: low-end microcontroller unit (MCU), high-end MCU, microprocessor unit (MPU), graphics processing unit (GPU), or neural processing unit (NPU). Understanding the advantages and disadvantages of each can help you choose the appropriate hardware for your edge AI needs.

You can read more about edge AI and take a practice quiz here: https://docs.edgeimpulse.com/docs/con...

Chapters:
0:00 Review of edge AI
0:37 Use case considerations
3:35 Interface considerations
4:29 Power constraints
5:08 Form factor considerations
5:30 Operating environments
6:06 Code portability
7:07 Buy vs. DIY
8:54 Low-end microcontrollers
10:41 High-end microcontrollers
11:39 Microprocessors
14:02 Graphics processing units (GPUs)
15:29 Neural processing units (NPUs)
16:43 Device specialization and cooperation