Research feasibility of solar power

The tRacket sensor currently depends on wired USB power. It would be massively beneficial if the sensor could be deployed long-term in locations where that isn't available; solar power seems like the most viable solution.

The sensor draws no more than 5V @ 100mA, and a quarter of that if WiFi is not running. Some ideas to look into:

* Solar panel options: USB panels? Solar harvesting circuits? How large of a panel do we need?
* Battery options: chemistry, capacity, size? A battery is necessary since the sensors run 24/7 and need energy when the sun is not out.
* Battery charger for storing excess solar energy into the battery
* PCB modifications: battery and/or solar panel connectors? Low-power microcontroller to minimize energy consumption when WiFi is not needed?
* Enclosure: Do we need a larger enclosure? How will the solar panel secure to it? More drilling for solar and battery connectors? Weatherproofing?
* How much will the solar-powered sensor cost?

I'll share more notes soon on the research I've done, but anyone is welcome to look into any of these points and share their findings here!