Hello,
I am attempting to simulate the LPS system. I am currently attempting to add noise to the state estimate of the simulated robot, but am unsure of where to add it in to reliably simulate error. As I see it, I have two options:
1) Add the noise when the position of the robot is computed
2) Add the noise in the distance difference measurement
I am using an extended Kalman filter like that in the Crazyflie.
I was also wondering how the measurement noise standard deviation (https://github.com/bitcraze/crazyflie-f ... 2Tag.c#L43) was chosen?
Thank you!
Thomas
Determination of Measurement Noise STD
-
- Beginner
- Posts: 23
- Joined: Thu Jun 28, 2018 12:00 am
Re: Determination of Measurement Noise STD
Hi!
The STD was measured as part of this paper:
Fusing ultra-wideband range measurements with accelerometers and rate gyroscopes for quadrocopter state estimation
Meuller et all (2015) ICRA (see here for pdf).
I believe they use a mocap system to measure it and they came to a 0.17 ns STD but that is now 0.15 ns for some reason (still pretty close)
If you would simulate the LPS system, I think you should go for option 2). The noise is supposed to be in the measurement anyway, so that will be closer to reality.
The STD was measured as part of this paper:
Fusing ultra-wideband range measurements with accelerometers and rate gyroscopes for quadrocopter state estimation
Meuller et all (2015) ICRA (see here for pdf).
I believe they use a mocap system to measure it and they came to a 0.17 ns STD but that is now 0.15 ns for some reason (still pretty close)
If you would simulate the LPS system, I think you should go for option 2). The noise is supposed to be in the measurement anyway, so that will be closer to reality.