in fast detection of forest fires

The design included the development of firmware for data analysis with energy constraints, support for over-the-air updates, IoT security, LoRaWAN implementation, and compliance with standards, the creation of proprietary links between Mesh Gateways, the integration of multiple forms of RF connectivity in frontier gateways and cloud platforms and analysis. All this was achieved with low-power equipment, with solar panels often in the shade, in the difficult RF environment of a forest, and with the robustness necessary to ensure long-term operation. And, above all, for fire sensors, all this was done at the lowest possible cost to allow for widespread diffusion.
 
A fundamental challenge was the design of the sensor hardware with very low power consumption. The system works in an environment where we have very little energy input, but we have to do something complex. We have to scan gas compositions and run machine learning software that works reliably, and we have to detect fires very quickly. These are conflicting objectives, so a challenge was to select the right components with very low power consumption.
 
At the same time, the price is crucial because hundreds of thousands if not millions of fire sensors will need to be placed in a forest for the system to work, and the price of each component at that scale determines the cost of the system.