undefined | Better HN

0 pointshmexx13y ago0 comments

Yeah, but the gyro/accelerometer could provide short movement measurements, no? IF you combine that with the distance measurement you get from the BT device, you should be able to figure out the direction to the BT device?

0 comments

zizzer13y ago

For simple dead reckoning you'd need to add a magnetometer into the mix as well, otherwise there's no way to know which direction the sensors are facing relative to anything else.

This would at least give you a compass heading that you could use when integrating the accelerometer data to give some kind of course over time.

It'd be pretty unreliable though and add a lot of cost.

hmexxOP13y ago

Still not sure why any of this would add cost. it all happens on the iphone which already has these sensors. If you can tell in which direction the iphone is moving thanks to gyro/magno/accelo AND you have a distance from BT device of reasonable resolution, you should be able to figure out the location of the device you are tracking with small movements of the iphone.

I'm not great at geometry but I figure it should be possible. Not sure about accuracy though.

zizzer13y ago

I was talking about adding the sensors to the bluetooth sticker device, that's why it'd add cost.

The sensors in your phone won't help to work out where the sticker is. The bluetooth receiver in it is built to detect signals from any direction, so even if you rotate your phone, that doesn't help to work out which direction the signals are arriving from.

1 more reply

krembo13y ago

Since each sticker is a transmitter and a receiver, isn't it possible to mount 3 stickers with known places around your house like GPS sattelites, which will calculate the cat's sticker (for example) distance from each of them and tell the exact location?

zizzer13y ago

Getting an exact location would be very tricky as the received signal strength can be affected by so many things (Thick walls, reflections, interference...).

I'd imagine you could get a rough location, but you'd need to go through some kind of calibration process first that involved going into different rooms/zones and telling the software where you were.

You might need extra reference sensors too. Ones that are in known locations that the system can use to scale the signal strength readings. That might help to combat some of the issues I mentioned above.

1 more reply

j / k navigate · click thread line to collapse

0 comments

zizzer13y ago

For simple dead reckoning you'd need to add a magnetometer into the mix as well, otherwise there's no way to know which direction the sensors are facing relative to anything else.

This would at least give you a compass heading that you could use when integrating the accelerometer data to give some kind of course over time.

It'd be pretty unreliable though and add a lot of cost.

hmexxOP13y ago

I'm not great at geometry but I figure it should be possible. Not sure about accuracy though.

zizzer13y ago

I was talking about adding the sensors to the bluetooth sticker device, that's why it'd add cost.

1 more reply

krembo13y ago

zizzer13y ago

Getting an exact location would be very tricky as the received signal strength can be affected by so many things (Thick walls, reflections, interference...).

I'd imagine you could get a rough location, but you'd need to go through some kind of calibration process first that involved going into different rooms/zones and telling the software where you were.

1 more reply

j / k navigate · click thread line to collapse