That way the plants do not have to be watered that often. The wooden sticks, which are sold at groceries for serving meat, are to keep the flowerpot from turning upside down, which might happen, because the mass center of the flowerpot is above the water level. Still water and mud is a biologically dangerous combination. The next prototype will probably have some sterilizing UV-LED-s at the bottom, probably mixed with some visible light LED-s for visual effects. May be the mud should be replaced with some hydroponic solution. The inspiration for the floating flowerpot comes from a documentary that describes the history of the Japanese semiconductor industry. The documentary describes a solution, where a similar mechanism was used for slowly pulling a rope, which in turn moved a radio wave based heating coil.
Once upon a time I created an electronic flower watering system as a school project. It was a microcontroller based system that was supposed to measure the humidity of the soil by measuring its electrical conductivity and, according to settings, which were entered through a keyboard, switched in a car windscreen solution pump to transfer water from a bucket to the flowerpot. The pumping side, the electronics and the software worked, I passed the course, but in practice the system was useless, because it was clumsy, it needed a separate bucket of water, which might not looked that aesthetic, and the worst part of the system was that the humidity sensing part did measure electrical conductivity of the soil, but the electrical conductivity of the soil does not necessarily reflect the humidity of the soil. It was really hard to figure out the conductivity level, on which the water pump should be switched on. A dirty safety measure was to use various heuristics, like: "do not water the flower, if it has been watered yesterday", "during one watering event the pump may be switched in only for N seconds to avoid flooding the floor, should the humidity sensor malfunction or fall out of the flowerpot", etc.
Actually, I think that the current flower-pot solution is literally ingenious, because:
- it solves the humidity measuring problem;
- it avoids all that complex electronics, which in turn means that it does not need any software development;
- it turns the un-aestetic water reservoir to a decoration element;
Too bad that the electronics could not be totally eliminated, but the solution to humidity measurement and automated watering is already a big win.
Update on 2014_11_19.
The plants at the photograph grew long roots and died. White spots, probably made of bacteria, developed on the soil.
Update on 2017_06_26.
After seeing the following image it struck me that in stead of trying to figure out, how to measure humidity directly, the whole flower pot might be placed ON A SCALE that monitors the weight of the whole flower pot.
The idea behind the watering algorithm might be that the starting value is when the plant has been properly watered and as the water evaporates through the plant's leaves, the flower pot as a whole gets lighter. If the flower pot becomes "too light", water is added to the pot automatically. The air based CO2 absorption related mass growth of the plant can be compensated by manually setting the weight counter to its starting position every time the water tank of the watering system is re-filled. There is no need for a water pump, because the system may use a gravity based system, like the intravenous therapy systems use. The watering electronics operates only a valve. Effects of an accidental flooding can be mitigated by placing the scale into a tub that is able to contain all of the stored water. It's OK for the open position of the valve to let through only a relatively narrow stream of water, because there's a lot of time available. To avoid triggering the watering action with vibration that spreads within the building structures and furniture, the weight of the flower pot is assessed by averaging the weight measurements over a period of a few minutes.