Cognitive IoT end device with deep learning power for motor monitoring: The solution optimizes the hardware of the product by integrating a dedicated embedded AI accelerator as the core processor while retaining other peripheral components for required functionalities. The hardware is designed, soldered, tested, and has now evolved to its second version. Additionally, software development, including the convolutionary neural network for the motor use case, is completed and deployed on the hardware system. For software development, a proper toolchain is acquired and implemented.
Furthermore, for power efficiency, both the hardware and software are optimized to ensure that the power consumption of the designed AI-based IoT device is theoretically small enough to run for two years when battery-powered.

