XCore Exchange

Problem: Real-world Monitoring

I would like someone to create an "environment" monitoring device. I know a customer who would like to purchase approximately 50 of these devices for a test run in a much bigger market. Anyone who is interested in designing and building this system should get in touch with me for more information.

This device is trained to recognise a particular physical input - assumed to be an analog waveform read from a sensor (it would be best to consider an audio input, but the design might be best left general) - and to record approximately timestamped instances of this input. It must be capable of reporting periodically (daily) on the number and pattern of inputs remotely - preferably via a mobile phone network. The device must be battery-operated and the aim is to last for at least a calendar year before a battery replacement is necessary.

The device should be easy to deploy: it should be placed it in the target environment and a button pressed to record the (audio) input to be detected (one or more times if necessary).

It doesn't have to use an XMOS chip, presumably.

leon_heller wrote:It doesn't have to use an XMOS chip, presumably.

No, not necessarily. I'm open to someone solving it with an XMOS chip or without.

By the way I didn't mention, you are not particularly space-constrained. The battery can be rather large if necessary.

jizzy_bear wrote:are you cost constrained?

In the specific market we're talking about, unit price of GBP40 = easy sale, GBP70 = tough sale, GBP100 = nearly impossible. Note that this will change over time. In 6 months, it's possible v2.0 can fetch GBP100 or so and still be purchased. But you need to prove business efficiency savings first!

Such a price seams pretty reachable, unless it needs to work in a tough environment (Like waterproof and such)

Berni wrote:Such a price seams pretty reachable, unless it needs to work in a tough environment (Like waterproof and such)

Really? Im interested in the battery technology you are thinking of in order to reach that price point given the specs?

I was thinking something like this might be suitable (although i am unsure of the price) the type XXI model could probably power a mobile device for about a year.

http://www.uboat.net/technical/batteries.htm

How about making it not need a huge battery. Use a low power MCU and use sleep mode as much as possible, so a couple of AAs should work fine. Its best to use a common battery instead of some internal lithium thing.

Berni wrote:How about making it not need a huge battery. Use a low power MCU and use sleep mode as much as possible, so a couple of AAs should work fine. Its best to use a common battery instead of some internal lithium thing.

If I understand correctly this device has to do audio recognition (of an unspecified type) 24/7 for a year and occasionally fire up a mobile phone. Will even the lowest power MCU available do 8000+ hours of DSP on a couple of AAs? That's abour 5000mAh over 8000 hours - 0.625 mW? That's not much.

You should add a solar panel to recharge where possible?