Comparing orientation data obtained using gyroscope and accelerometer

Question

Recently, I have been playing around with a gyroscope and accelerometer in the hopes of building a quadcopter. I have plotted some of the data the two sensors and I am noticing that the pitch from the accelerator lags significantly behind the gyroscopic pitch.

From some research, I understand that these two are not going to equivalent values due to the mechanisms of measurement, however is it common to have such large discrepancies (i.e a difference of 20 degrees or higher) between the two sensors or is this an issue with my conversion?

I have listed the graph and source code here. Any help would be appreciated!

Sincerely, Anand

Answer 1

Normally the values from gyroscope will drift overtime. It is natural and depends on the parameters of your code. It could drift a few degrees per seconds and the data will be far way from the accelerometer in a few seconds.

I have seen that you have applied the complementary filter, so you don't need to worry about the single value from the gyros since you combine them with the acceleromter and everything seems fine.

The gyros are good for short time response, in long time response the values get pretty bad as they drift a lot.

On the other hand, accelerometers are good for long time response, since short time response is noisy because of gravitational influence.

That's why you need a filter to combine them.

Wrapping up, it is normal to have this behavior in gyros. If you want to know more about the effect behind this sensor, try this reading.