boakes.org

On a recent train journey I overheard a conversation about someone who’d enjoyed the first class wine so much on a recent intercity journey that they had difficulty staying awake so as not to miss their destination. Staying awake is a problem for many commuters and for those whose destination is not the terminating station, it can be both costly and time consuming. An alarm clock is the obvious tool of choice, to solve the problem, but it’s not convenient. Helpfully some phones include alarm clock applications, but aside from the fact that they can ring a bell at a predefined time, they’re not very useful because what’s important is getting off at the right station, not at a specific time.

Some commuters always catch the same train, at the same time, so a time based alarm might do the job, but for anyone with more flexible train usage, and for the times when the train is running late, or cancelled, a simple alarm clock doesn’t cut the mustard. So how about an alarm system that you could configure to go off every time you come within a set distance of a specific point? I’m sure it’s something my GPS phone could no doubt do quite successfully, and it would solve the problem of arriving at the same train station at different times.

Firstly, it’s not the kind of application you could leave running on the phone and just forget, because it would fire every time you got near the station. If you happen to live near the station the alarm would be going off at all times. This could be solved if the alarm were based on a series of way-points. i.e. If you pass near points a, b and c in sequence then the alarm should trigger. This solves the problem of the alarm going off when it’s not necessary. The main problem however is that the application still cannot be left running and forgotten because of the battery usage - the Nokia N95 for example, has a good battery and during normal phone usage it lasts well, but the circuitry that does the GPS appears to have a far more significant draw on power than the rest of the device, killing the phone in a matter of hours.

There is, however, a suitable alternative to GPS. Rather than base the alarm on GPS locations, it can be based on cellphone mast IDs. As the commuter moves through the country, the phone detects, and negotiates with different telephone masts. If the alarm system were to be configured to work with masts, then it could work quite successfully without the need for any GPS circuitry to be used.

It’s certainly a less accurate positioning solution than GPS, so the multi-stage trigger for the alarm would be a necessity, but this could be a fairly low-power service because the application would merely have to be woken up whenever a new cell is discovered by the phone, checked for relevance (i.e. does this new cell match an alarm? are any alarms satisfied for triggering?). It’s still not perfect though. What’s needed is a minimal battery solution for that we need to remove the application altogether.

The perfect solution, would remove the need for the phone based application altogether. If the phone company could monitor the phone as it travels between cells, and send an SMS to the phone when the the trigger criteria are met. This solution would work for all phones, even old ones which can’t have applications added to them. No commuter need ever miss their stop again.