Edge AI air purifier that predicts pollution spikes before they happen.

A home appliance brand wanting to differentiate their purifier line with AI-driven preemptive air quality control.
Building a cost-effective air purifier controller that can learn pollution patterns (cooking hours, traffic spikes) and pre-activate filtration before air quality degrades.
The Rockchip RK3308 quad-core Cortex-A35 SoC was chosen for its Linux capability and rich peripheral set. It reads from a Sensirion SPS30 PM sensor, an SGP40 VOC sensor, and a BME280 temperature/humidity sensor. A PWM-controlled Nidec brushless DC fan provides airflow.
We deployed Buildroot Linux with a custom Python/Node.js edge application. A lightweight LSTM model trained on 3 months of household air quality data runs on the RK3308's Cortex-A35 cores. The model predicts PM2.5 levels 30 minutes ahead with 92% accuracy. When a spike is predicted, the fan ramps up preemptively. The device also publishes hourly aggregates to an MQTT dashboard.
User exposure to PM2.5 reduced by 35% compared to reactive auto-mode. Filter life extended 20% through smarter fan scheduling. Total BOM under $45.
Let's discuss how ChipTalk can deliver results for your next project.