How long does the battery need to last? Does the Arduino need to be able to send text messages DIRECTLY via GSM? Or can it send through some other fairly local network device? Must it be completely wireless (usually the implication with a battery, but not always), or can it plug into ethernet? Or even better, a USB port on something else close by (unlikely, but gotta ask)?
Are you talking about doing it "as small as possible" but still use fairly "plug and play" modules like existing Arduino boards and shields, or are you willing to do wiring and/or make custom boards?
What kind of "tilt sensor"? Do you need to know if it moves anywhere from one degree to 90? Or just know if it tilts past a certain point? Do you need to know how much it tilted via the text message, or just that it tilted at all?
You just can't expect people to spec your project for you.
Realistically your are governed by the size of two components - the battery and the gsm modem. Simplest way is Uno, GSM Shield, tilt switch (I'm assuming a mercury type tilt switch in the absence of information) and battery.
These are the smallest Arduino compatible devices that I'm aware of. They include a 2.4GHz wireless transceiver. I use them in N-Gauge model trains for radio control.
Thanks for the replies and apologies if I didn't give enough info.
My goal is to create "something" that I can store it under my motorcycle seat. It will sense if the bike is being fiddled about (moved/tilted effectively) and will send me a text message on my mobile phone.
Battery wise, I guess I could plug it in on the motorcycle battery, so I guess everything else is priority.
Size is a big issue, as it must be as small as it can get.
Does it make more sense now what I am trying to achieve?
Assuming you want to continue down the GSM modem route, where would the aerial go? Even on modern bikes there's a lot of metal around there to block the signal.
For this scenario there aren't too many options, but I'm just trying to think of the most vulnerable parts. A traditional aerial could also be broken off, rendering the system useless. I've not had a play with the newer 'flat' aerials yet so that may be one avenue to explore.
Is there a reason you aren't going for a straight forward bike alarm?
A bike alarm most usually does not have a tracker and it won't notify you with texts if the bike is tampered with.
My bike was stolen with an alarm on it and a number of chains and other security items on it. They just picked it up, put it in a van and it was gone.
The trackers that offer such functionality are extremely expensive and have a monthly subscription to their service making it not affordable. I can't justify paying £700-800 (first year and subscription) and then another £200-300 a year for such service.
If the bike is gone... having it tracked down doesn't really make a difference since by the time you get to it, it's already been stripped down and pretty much a write-off.
That's why I thought I should head down this path and try to make one of my own.
Having said that... would a Leonardo board, a tilt sensor and a gsm shield would suffice?