Ok m so now i am doing an indoor positioning system, i hook up some accelerometer and gyroscope to a user and use it to determine the position the user is at. Using accelerometer and gyroscope i am able to find out the yaw pitch and roll of the imu .Under 2d circumstances i am able to determine the position of the user based on imu and magnetometer, however i have no idea how to determine the user position when he or she is climbing stairs (3d) any idea?

 

Instead of trying to control using an accelerometer directly, and if you're willing to dig into the math, the most common way to do this is to combine a barometer and accelerometer's measurements using a Kalman filter. This will give you a smooth estimate of the altitude and velocity.

 

Ok m so now i am doing an indoor positioning system, i hook up some accelerometer and gyroscope to a user and use it to determine the position the user is at. Using accelerometer and gyroscope i am able to find out the yaw pitch and roll of the imu .Under 2d circumstances i am able to determine the position of the user based on imu and magnetometer, however i have no idea how to determine the user position when he or she is climbing stairs (3d) any idea?

This is a very advanced project.

What accelerometer data do you have to show?

If you examine the vertical axis accel data you should find for each step of the stairs there is a definiable hump on the Z axis data.

If you show us your datalog of the axis data over time we can probably help more.

 

This is a very advanced project.

So true. The latest iPhone 5S with its M7 motion analysis chip is probably state-of-the-art (for affordable commercial products, anyway) and it is by no means perfect for the task. Yet. Rumors abound regarding what the iPhone 6 will do.

 

This is a very advanced project.

What accelerometer data do you have to show?

If you examine the vertical axis accel data you should find for each step of the stairs there is a definiable hump on the Z axis data.

If you show us your datalog of the axis data over time we can probably help more.

I didnt just use the accelerometer data alone, i used the yaw pitch roll data which is processed from the gyroscope and the accelerometer.

And dueing these few night i was thinking, since when moving upstairs our body will somesort like move upwards, is there any sensor that i can soemsort like place on my chest so that it can detect the change in angle when my body move upwards, if like that i can noe when the user is climbing the stairs. And also by placing the sensor on the chest , u will get a much more accurate result compare to putting it on the leg, because u seldom to a big movement with ur chest, is my suggestion implementable?

 

I do not think angle of the torso alone would be very sensitive. Some people bend minimally going up or down stairs.

Have you considered a measure of the angle of the thigh (femur) to the torso? Unless someone has an awfully odd walking gate, that angle should be more acute (smaller) going up or down stairs. Of course running and maybe some disabilities may affect that relationship.

Do you have to rule out sitting?

John

 

The FitBit (accelerometer worn on the wrist) can tell when you are climbing stairs. It's all about the algorithm.

 

A torso mounted accelerometer will give Z axis data humps for each step, humans don't move their torso linearly up stairs.

The questions the OP is asking seem to indicate that he has not done much testing.

 

I do not think angle of the torso alone would be very sensitive. Some people bend minimally going up or down stairs.

Have you considered a measure of the angle of the thigh (femur) to the torso? Unless someone has an awfully odd walking gate, that angle should be more acute (smaller) going up or down stairs. Of course running and maybe some disabilities may affect that relationship.

Do you have to rule out sitting?

John

I know this is a very advanced project, i am thinking of a simple way to do this without going through many mathematical equation. For 2d i am placing the imu on my hip and i uses the length of the leg and the change in angle to calculate the distance that i have travel and uses magnetometer to calculate the direction it is heading, i haven try out but at least i have an idea of how to do it, but when it comes to 3d , if i get the same data of the imu that i hooked on my leg,i have no idea how to differentiate whether i am runnning or whether i am climbing the stairs

 

A torso mounted accelerometer will give Z axis data humps for each step, humans don't move their torso linearly up stairs.

The questions the OP is asking seem to indicate that he has not done much testing.

Eh so means actually by just using the data from the accelerometer, it is implementable? I dont need to use yaw roll and pitch?

 

... i am thinking of a simple way to do this without going through many mathematical equation...

Good luck with that. A lot of well-funded smart people are also working on this, and I'm quite sure they are using plenty of math.