Is it possible to calculate precise speed with LSM6DS3 accelerometer?

[VAvag.1]

..

[Peter BENSCH]

Well, an accelerometer can only measure acceleration in different axes. If it moves at a constant speed in a constant direction (linear vector), the acceleration is zero. So you can only measure changes in velocity and calculate a change in position from this by integration, for example, whereby drift effects must of course be taken into account.

But a velocity (speed) cannot be derived from this.

Does it answer your question?

Regards

/Peter

[VAvag.1]

Hi Peter. Thank You for response.

My aim is to calculate the speed in one direction (x-axis). Movement begins from 0 speed and ends with 0 speed, There are feu speeds (including intermediate 0 speed) that I must distinguish. As You mention, if acceleration is 0, the speed is constant, and it may be not the same as a the beginning. But if the sum of all sampled acceleration values (positive & negative) for particular interval is 0, then we can say that the speed at the end of that interval is same as at the beginning (particularly 0 beginning 0 end). In real life, when I take samples and calculate speed, the graphic have a slope (positive or negative), and it never ends at the same speed as at the beginning, even I take into account Zero-g level. Are some correcting algorithms for it? Sending You 2 graphics, 1-st is simple movement in 2 directions, an the 2-nd is real movement. In both cases real movement stars and stops at 0 speed.