Legacy Forum: Preserving Nearly 20 Years of Community History - A Time Capsule of Discussions, Memories, and Shared Experiences.

3d Accelerometers - Will they do position sensing??

Anything that doesn't fit our other forums goes here.
7 postsPage 1 of 1
7 postsPage 1 of 1

3d Accelerometers - Will they do position sensing??

Post by hs92 » Wed Dec 02, 2009 10:39 am

Post by hs92
Wed Dec 02, 2009 10:39 am

does anyone know much about the 3D accelerometers that are on here for sale?

I am looking to use one for a positioning system possibly using one of the arduino boards.

I need to be able to work out where my device is at a given time, eg, its distance travelled since last calculation in all three dimensions.

If anyone can help me out i would be very grateful and if you need to know any more info just ask away,

Thanks,

Henry
does anyone know much about the 3D accelerometers that are on here for sale?

I am looking to use one for a positioning system possibly using one of the arduino boards.

I need to be able to work out where my device is at a given time, eg, its distance travelled since last calculation in all three dimensions.

If anyone can help me out i would be very grateful and if you need to know any more info just ask away,

Thanks,

Henry
hs92
Newbie
Newbie
Posts: 4
Joined: Wed Dec 02, 2009 10:36 am

Post by i-Bot » Wed Dec 02, 2009 2:39 pm

Post by i-Bot
Wed Dec 02, 2009 2:39 pm

Accelerometer measure acceleration exactly as they say. To get position you have to integrate the signal output twice. Integration intoduces large errors which accumulate to make the measurement of position almost useless. Accelerometers also detect the acceleration due to gravity, so are best for determining orientation about a vertical axis, which is why they are used so much in robotics.

Measuring distance travelled is better done with a direct position measurement. This depends on the scale (millimetres, metres, kilometers) and the relationship of the device to its surroundings ( wheels, known environment, fixed or autonomous). Different techniques can then be used including encoders, cameras, gps, laser, IR.

So I don't think accelerometers will do it for you, but there is usually a way.
Accelerometer measure acceleration exactly as they say. To get position you have to integrate the signal output twice. Integration intoduces large errors which accumulate to make the measurement of position almost useless. Accelerometers also detect the acceleration due to gravity, so are best for determining orientation about a vertical axis, which is why they are used so much in robotics.

Measuring distance travelled is better done with a direct position measurement. This depends on the scale (millimetres, metres, kilometers) and the relationship of the device to its surroundings ( wheels, known environment, fixed or autonomous). Different techniques can then be used including encoders, cameras, gps, laser, IR.

So I don't think accelerometers will do it for you, but there is usually a way.
i-Bot
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 1142
Joined: Wed May 17, 2006 1:00 am

Post by hs92 » Wed Dec 02, 2009 4:11 pm

Post by hs92
Wed Dec 02, 2009 4:11 pm

Thanks for the reply. I am looking at a device that will be travelling down a pipe in a sealed unit whilst a medium is flowing. So the device will need to know where it is whilst travelling down the pipe.

From what you are saying, an accelerometer will tell me the orientation of my device, in terms of x-y-z coords.

But to measure the position in the pipe at a given time will require something else...Could you possibly suggest anything that may work?

Thanks.
Thanks for the reply. I am looking at a device that will be travelling down a pipe in a sealed unit whilst a medium is flowing. So the device will need to know where it is whilst travelling down the pipe.

From what you are saying, an accelerometer will tell me the orientation of my device, in terms of x-y-z coords.

But to measure the position in the pipe at a given time will require something else...Could you possibly suggest anything that may work?

Thanks.
hs92
Newbie
Newbie
Posts: 4
Joined: Wed Dec 02, 2009 10:36 am

Post by hs92 » Wed Dec 02, 2009 4:42 pm

Post by hs92
Wed Dec 02, 2009 4:42 pm

Could you not use the acceleration in the direction of travel, over a given time to calculate the distance travelled?

Cheers,
Could you not use the acceleration in the direction of travel, over a given time to calculate the distance travelled?

Cheers,
hs92
Newbie
Newbie
Posts: 4
Joined: Wed Dec 02, 2009 10:36 am

Post by SK » Thu Dec 03, 2009 1:14 am

Post by SK
Thu Dec 03, 2009 1:14 am

hs92 wrote:Could you not use the acceleration in the direction of travel, over a given time to calculate the distance travelled?

Cheers,

Well you could try that, but I´d strongly suggest you don´t get your hope too high that it works for more than a very short time (seconds). After all, you´d be still integrating noise twice.
Of course, one could use (Kalman) filtering for fusing data from the accelerometers with some other measurement device that might give less frequent but more accurate updates (RFIDs spring to my mind for example).
It´s probably a good idea to read up on Kalman Filters, INS/IMU systems and low cos inertial sensors before trying anything :)
hs92 wrote:Could you not use the acceleration in the direction of travel, over a given time to calculate the distance travelled?

Cheers,

Well you could try that, but I´d strongly suggest you don´t get your hope too high that it works for more than a very short time (seconds). After all, you´d be still integrating noise twice.
Of course, one could use (Kalman) filtering for fusing data from the accelerometers with some other measurement device that might give less frequent but more accurate updates (RFIDs spring to my mind for example).
It´s probably a good idea to read up on Kalman Filters, INS/IMU systems and low cos inertial sensors before trying anything :)
SK
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 105
Joined: Mon Jun 26, 2006 1:00 am
Location: near Frankfurt/Main, Germany

Post by hs92 » Thu Dec 03, 2009 2:19 pm

Post by hs92
Thu Dec 03, 2009 2:19 pm

Thanks for the reply.

So it would be best to use another means for the distance travelled.

If i want to know the position in the pipe of the device at any one time in terns of x-y coords, ie height in pipe and position across the pipe, would accelerometers give me what i want or would i have to use some other means?

Cheers.
Thanks for the reply.

So it would be best to use another means for the distance travelled.

If i want to know the position in the pipe of the device at any one time in terns of x-y coords, ie height in pipe and position across the pipe, would accelerometers give me what i want or would i have to use some other means?

Cheers.
hs92
Newbie
Newbie
Posts: 4
Joined: Wed Dec 02, 2009 10:36 am

Post by SK » Fri Dec 04, 2009 8:53 am

Post by SK
Fri Dec 04, 2009 8:53 am

Well I haven´t tried such an application myself, so I can only state my other experience with low cost accelerometers which basically says "don´t use them for unaided position estimation for timespans larger than a few seconds max".

A potentially useful article dealing with 1D accelerometer filtering is this:
http://www.gamasutra.com/view/feature/1 ... alman_.php
Well I haven´t tried such an application myself, so I can only state my other experience with low cost accelerometers which basically says "don´t use them for unaided position estimation for timespans larger than a few seconds max".

A potentially useful article dealing with 1D accelerometer filtering is this:
http://www.gamasutra.com/view/feature/1 ... alman_.php
SK
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 105
Joined: Mon Jun 26, 2006 1:00 am
Location: near Frankfurt/Main, Germany


7 postsPage 1 of 1
7 postsPage 1 of 1