Rotation angle from angular rate

In summary, Chathuri is using an inertial sensor from Analog Devices to obtain rotation angle. They have integrated the angular rate to calculate the rotation angle in real time, but have noticed a slight increase or decrease in rotation angle when the sensor is held stationary. This is believed to be caused by a drift from the integration. Chathuri is seeking advice on how to fix this issue and has mentioned using a separate sensor to detect when the device is level or modeling the drift and applying a correction. Cleonis is unable to provide further assistance.
  • #1
chathuri
3
0
Hi,

I am using an inertial sensor to obtain rotation angle. However the sensor only gives the angular rate and acceleration. Therefore I integrated the angular rate to obtain the rotation angle in real time. However, I noticed that if I keep the sensor stationary, the rotation angle starts to increase or decrease slightly but continuously. I believe this has somethin to with a drift caused by the integration.

Would anyone know how I could fix this issue. I am very knew to this and was hoping you could help me out.

Thankyou!

Chathuri
 
Physics news on Phys.org
  • #2
It's an intrinsic problem of all rate sensors. (It's why 727s have a window above the cockpit to see the stars)
Best option is to be able to reset to a fixed point periodically - eg by having a separate sensor that detects when it's level.
Second best is to model the drift and apply a correction - but the drift is likely to be temperature dependant and may differ from one unit to the next.
 
  • #3
Thanks for your reply. But I need to clarify. Do you mean I should check if the sensor is stationary (probably see if the acceleration is close to 0) and then keep the previous value of the rotation angle without integration?

If acceleration <0.01 then theta = theta_prev

Thankyou!

Chathuri
 
  • #4
If you have an acceleration sensor then when it reads zero you don't know that it is in any particular angle.
To o this accurately you need an external reference, such as a separate level sensor that triggers when the device is exactly horizontal and use this to reset your angle integrator. Or you might be able to get a reset when, for instance, a shaft passes some particular point.
 
  • #5
chathuri said:
Hi,
I am using an inertial sensor to obtain rotation angle. However the sensor only gives the angular rate and acceleration. Therefore I integrated the angular rate to obtain the rotation angle in real time. However, I noticed that if I keep the sensor stationary, the rotation angle starts to increase or decrease slightly but continuously. I believe this has somethin to with a drift caused by the integration.

You have not described in any way how you have implemented the integration. What kind of device performs the integration? Is the device such that a systematic bias may occur? If so, how large a systematic bias? Can it account for the drift?

You are not providing enough clues to give me something to work with.What do you mean with: 'When I hold the sensor stationary'? For instance, if a rotation sensor is stationary with respect to the Earth, then it will register the Earth's rotation rate. (Depending on the type of rotation sensor the sensor's orientation with respect to the Earth's axis will also be a factor in how much rotation is registered.)

High performance rotation sensors are easily sensitive enough to register the Earth's rotation. Effectively, the prime judge of 'being stationary' is the rotation sensor itself. If it registers zero rotation rate then it's stationary.

Cleonis
 
Last edited:
  • #6
Hi

I am sorry for not explaining myself clearly. The sensor that I am using is ADIS16350 from Analog Devices. Before taking reading I use a command to correct the gyroscope bias. Then I am using the following equations to calculate the rotation angle (theta) from the angular rate (w).

theta(k+1) = (w(k+1) - w(0))*t + theta(k);

Here t=0.01 as the sampling rate is set to 100Hz. The w(0) is zero frequency offset of the rate gyro and is determined as the average of the angular rate for the initial 49 samples. 'k' represents the sample number.

When the sensor is held with no motion (rate gyros readings are close to zero but not zero) the drift occurs. I believ this is because the rate gyro gives a slight reading.

Is there a conventional method to deal with the drift? I am sorry if I am asking obvious questions. It is just that I am very new to dealing with sensors.

Thank you!

Chathuri
 
  • #7
I apologize for doubting the thoroughness of your preparations. Clearly you know as much as I do. I have to confess I'm not better equipped than you are to find where the drift is entering the system. Good luck with your efforts.

Cleonis
 

What is a rotation angle?

A rotation angle is a measure of the amount of rotation an object has undergone around a particular axis. It is usually measured in degrees or radians.

What is angular rate?

Angular rate is the rate at which an object is rotating around a particular axis. It is usually measured in degrees or radians per unit of time.

How are rotation angle and angular rate related?

The rotation angle and angular rate are directly proportional to each other. This means that as the angular rate increases, the rotation angle also increases, and vice versa.

What is the formula for calculating rotation angle from angular rate?

The formula for calculating rotation angle from angular rate is: rotation angle = angular rate x time. This means that the rotation angle is equal to the angular rate multiplied by the amount of time the object has been rotating.

Why is understanding rotation angle from angular rate important in science?

Understanding rotation angle from angular rate is important in science because it helps us to analyze and predict the motion of rotating objects. This knowledge is crucial in various fields such as physics, engineering, and astronomy.

Similar threads

Replies
2
Views
723
Replies
2
Views
646
  • Mechanical Engineering
Replies
3
Views
361
Replies
4
Views
695
Replies
22
Views
1K
  • Mechanics
Replies
9
Views
4K
Replies
3
Views
1K
Replies
6
Views
3K
Replies
1
Views
1K
Back
Top