On Monday we discussed the Pine Mountain Settlement school and how the structures will affect the design of our sensors. Towards the end of the discussion, the idea of tapping into their electrical system was brought up. One of my colleagues suggested that we just use batteries to power the sensors, and use wifi to log the data to an external dashboard/ google spread sheet. The theory behind this conjecture is that it will be simpler, and easier to manage. All of the buildings that the sensors will go in have wifi. However, not all of the buildings have access to electric outlets.
I don’t necessarily agree that the use of batteries is the simplest method. This is because in order to get a battery to last a long time, you must create a power savings method. For example, the majority of the time your machine should be sleeping in order to conserve power. It should wake up only at certain times (and only a certain number of times per day) to take data readings and log them. This can be a very tricky situation. The solution will involve external libraries (which are not difficult to use, but will require some insight) and it may also involve a real time clock (in order to wake up the sensor). Due to the fact that not all of the buildings have electrical outlets, we will be faced with solving this issue regardless of if we choose to deploy batteries in all of our sensors. Once we solve it for one sensor, it is as good as solved for the rest of the sensors because they will require only slight modifications (i.e. How often they wake up). Therefore, I do believe that the battery solution is a valid one, though it may not be as simple as it seems.
The battery solution also allows us to avoid tapping into Pine Mountain’s electrical system. If we went that route and something were to go wrong, then it would possible that we could have contributed to the destruction of a building. However, the battery solution poses other problems as well. How long with the batteries last? When will they need to be changed? Who will change them? Or can they somehow be externally charged? Can you tell how long the battery has left to live? I do believe that our sensor will be able to communicate about how much battery power is left. I do not know the answers to the other questions. If the batteries do die after (lets pretend) one year, I think it may be kind of ridiculous for someone to have to drive all the way to Pine Mtn to change the batteries. Unless they are looking for an excuse to go (haha). However, this means that changing of the batteries should be as uncomplicated and user friendly as possible (of course unless we have charging stations).
However, to sum it up in a sentence, I do believe that the batteries are a better choice than tapping into the electrical system. Although, that would make our sensors more permanent. There are of course pros and cons to each. The pros to the electrical system would be no issue of conserving power, and we could take as many readings per day as we like. There would be no batteries to change. However, like what was said in class, we are not licensed electricians. Messing around with the electrical systems in old buildings sounds rather sketchy to me. If we can find a feasible solution that avoids that, I say we forge ahead.
The wifi seems like a simpler argument for me than the batteries. This is because it seems to make sense. There is wifi all throughout each of the buildings that there could potentially be sensors in and transmitting data over wifi sounds like a solid plan. I think the other options that we have are using radio dongles? I am not quite sure how this would work. However, a quick google search turned up this. My first question is : Holy cow are those prices correct? $75 to $85 dollars for wifi shields/ radios? No way that is right. Anyways, they have some of the core concepts correct. Radios can be used anywhere to talk to another radio, wifi only works on a wifi network.
Using wifi means that we do not have to fuss with having an extra radio to be the receiver. However, it brings up several questions. Such as, we want these sensors to work on the coldest of mornings (as well as the hottest) and be reliable in extreme weather conditions. Well, when the weather starts to get unsightly, sometimes the power goes out. No power means no wifi. No wifi would mean no sensors. Although, I am sure that in this situation, the people of Pine Mountain would be checking their buildings anyways. But who am I to say? It would be nice to be able to connect to your database and check the temperatures of all of your buildings without having to step foot out in the weather (if unnecessary). However, these are just my own ramblings.
In general wifi seems like the way to go but we should consider using radios when we have a conversation about reliability. Would radios be more reliable than wifi dongles? What are the pros and cons of each? In the end, it seems that if the buildings do all have wifi, we should take advantage of that.