Using scripts to process data from the FlightGear simulator #121

brightproject · 2024-09-07T13:22:49Z

brightproject
Sep 7, 2024

I receive acceleration and angular velocity data from the FlightGear simulator via serial port.
Firstly, I couldn't get the WSL scripts to work in Windows.
Python3 simply didn't want to run the setup as described here.

https://github.com/Mayitzin/ahrs?tab=readme-ov-file#installation

 python -m pip install .
Command 'python' not found, did you mean:
  command 'python3' from deb python3
  command 'python' from deb python-is-python3

That's why I'm just trying to run the Python code via PyCharm, but I haven't found a single example of working with the AHRS library in the folders themselves.
How can I insert data into it to determine orientation?
For example, I insert acceleration and angular velocity into the Mahony filter, either from the MEMS sensor or from the *csv file.
I don't know how to use the AHRS code in Python, although I have sufficient experience programming in C++.
This is not a complaint, it's just that the author of the library didn't really describe working with the library, only general phrases.
Also, the very first issue on GitHub says that the author will help everyone who is interested in the AHRS library, but in fact, most of the questions here are unanswered, which suggests that the project has been abandoned.
Please change me for such a tone, but until I can figure out this code, it will remain negative (tone).
Sorry!

brightproject · 2024-09-07T13:41:15Z

brightproject
Sep 7, 2024
Author

When I started writing this discussion, I knew that sooner or later I would figure out a little bit about the algorithms.
I found a little more information in another repository

https://github.com/FlyTheThings/ahrs-1

I managed to run some examples right in the console.
As far as I understand, this way you can get a rotating matrix of cosines:

0 replies

brightproject · 2024-09-08T06:54:03Z

brightproject
Sep 8, 2024
Author

I changed the example code a bit new_test_filters.py
But nothing is displayed on the graphs, actually I asked about such examples in this discussion... it seems like there is a lot of information, but at the same time there is nothing specific about the project.

import numpy as np
import ahrs
import matplotlib.pyplot as plt

RAD2DEG = ahrs.common.RAD2DEG
DEG2RAD = ahrs.common.DEG2RAD

class Test_Filter:
    def __init__(self, **kwargs):
        self.file = 'repoIMU.csv'
        self.data = ahrs.utils.io.load(self.file)


    def check_integrity(self, Q):
        sz = Q.shape
        qts_ok = not np.allclose(np.sum(Q, axis=0), sz[0]*np.array([1., 0., 0., 0.]))
        qnm_ok = np.allclose(np.linalg.norm(Q, axis=1).mean(), 1.0)
        return qts_ok and qnm_ok

    def allocate_arrays(self):
        self.Q = np.zeros((self.data.num_samples, 4))
        self.Q[:, 0] = 1.0

    def plot_quaternions(self, title):
        time = np.arange(self.data.num_samples)
        plt.figure(figsize=(12, 8))
        plt.subplot(2, 2, 1)
        plt.plot(time, self.Q[:, 0], label='q0')
        plt.title(title + " - q0")
        plt.xlabel('Time')
        plt.ylabel('q0')
        plt.grid()

        plt.subplot(2, 2, 2)
        plt.plot(time, self.Q[:, 1], label='q1', color='orange')
        plt.title(title + " - q1")
        plt.xlabel('Time')
        plt.ylabel('q1')
        plt.grid()

        plt.subplot(2, 2, 3)
        plt.plot(time, self.Q[:, 2], label='q2', color='green')
        plt.title(title + " - q2")
        plt.xlabel('Time')
        plt.ylabel('q2')
        plt.grid()

        plt.subplot(2, 2, 4)
        plt.plot(time, self.Q[:, 3], label='q3', color='red')
        plt.title(title + " - q3")
        plt.xlabel('Time')
        plt.ylabel('q3')
        plt.grid()

        plt.tight_layout()
        plt.show()

    def fourati(self):
        self.allocate_arrays()
        fourati = ahrs.filters.Fourati()
        for t in range(1, self.data.num_samples):
            self.Q[t] = fourati.update(DEG2RAD*self.data.gyr[t], self.data.acc[t], self.data.mag[t], self.Q[t-1])
        if self.check_integrity(self.Q):
            self.plot_quaternions("Fourati Filter")

    def ekf(self):
        self.allocate_arrays()
        ekf = ahrs.filters.EKF()
        for t in range(1, self.data.num_samples):
            self.Q[t] = ekf.update(DEG2RAD*self.data.gyr[t], self.data.acc[t], self.data.mag[t], self.Q[t-1])
        if self.check_integrity(self.Q):
            self.plot_quaternions("EKF Filter")

    def mahony(self):
        self.allocate_arrays()
        mahony = ahrs.filters.Mahony()
        for t in range(1, self.data.num_samples):
            self.Q[t] = mahony.updateMARG(DEG2RAD*self.data.gyr[t], self.data.acc[t], self.data.mag[t], self.Q[t-1])
        if self.check_integrity(self.Q):
            self.plot_quaternions("Mahony Filter")

    def madgwick(self):
        self.allocate_arrays()
        madgwick = ahrs.filters.Madgwick()
        for t in range(1, self.data.num_samples):
            self.Q[t] = madgwick.updateMARG(DEG2RAD*self.data.gyr[t], self.data.acc[t], self.data.mag[t], self.Q[t-1])
        if self.check_integrity(self.Q):
            self.plot_quaternions("Madgwick Filter")

I also tried to simply read data from a file using this code:

import numpy as np

# Load data from file
data = np.genfromtxt('repoIMU.csv', delimiter=';', skip_header=2)

# Extract time and quaternions
time = data[:, 0]
quaternions = data[:, 1:5]

# Extract accelerometer, gyroscope, and magnetometer data
acceleration = data[:, 5:8]
gyroscope = data[:, 8:11]
magnetometer = data[:, 11:14]

# Print data for verification
print("Time:", time)
print("Quaternions:", quaternions)
print("Acceleration:", acceleration)
print("Gyroscope:", gyroscope)
print("Magnetometer:", magnetometer)

The result in the console, after command python3 new_test_filters.py is as follows:

Time: [8.000e-02 9.000e-02 1.000e-01 ... 8.998e+01 8.999e+01 9.000e+01]
Quaternions: [[ 1.0000e+00 0.0000e+00 1.6232e-35 0.0000e+00]
[ 1.0000e+00 1.4021e-06 1.8673e-06 -2.9199e-07]
[ 1.0000e+00 2.8249e-06 3.7031e-06 -2.5293e-07]
...
[ 1.6123e-01 9.8625e-01 -3.6266e-02 -1.3439e-03]
[ 1.5934e-01 9.8656e-01 -3.6239e-02 -1.3452e-03]
[ 1.5747e-01 9.8686e-01 -3.6211e-02 -1.3442e-03]]
Acceleration: [[-0.071594 0.21157 9.7958 ]
[-0.071702 0.21878 9.7911 ]
[-0.081245 0.20916 9.7764 ]
...
[-0.023582 2.559 -9.1003 ]
[-0.051928 2.5768 -8.879 ]
[-0.07637 2.61 -8.8695 ]]
Gyroscope: [[ 2.3140e-03 -6.3400e-03 1.3220e-03]
[ 5.3300e-04 -7.2870e-03 -9.5700e-04]
[ 9.3420e-03 -4.2400e-04 3.5850e-03]
...
[ 3.6318e-01 -1.2522e-02 -3.7300e-04]
[ 3.7517e-01 -8.7670e-03 -1.5938e-02]
[ 3.7059e-01 -2.6818e-02 -1.7078e-02]]
Magnetometer: [[-0.33533 0.19856 -0.88708]
[-0.33402 0.20025 -0.88472]
[-0.33514 0.19707 -0.88613]
...
[-0.35902 -0.45602 0.83672]
[-0.35677 -0.45603 0.83675]
[-0.35742 -0.4499 0.84045]]

I'll try, there are no really working examples to understand this zoo of functions.
For a newbie in Python and in general in the topic of AHRS, this is very, very bad.

0 replies

Mayitzin · 2024-09-09T08:10:04Z

Mayitzin
Sep 9, 2024
Maintainer

Hi! Sorry about the lack of clarity, especially for beginners, in the repository. You have raised a good point, and I think I can offer a good solution for it.

I will add a Jupyter Notebook with the most common uses of the functions and tools included in the repository. This way I can explain the structure of the package and its possible applications.

I hope I can improve the readability for everyone, and its use at all levels.

I'll write back here to ask for feedback about the notebook.

Mario.

1 reply

brightproject Sep 9, 2024
Author

I'll write back here to ask for feedback about the notebook.

@Mayitzin I just like to be indignant, that's my nature.
Perhaps, in the process of sorting out the code, I will be able to write several examples of working with the library.
Ideally, I want to get a solution to place the data received from the flight simulator into the code algorithms and receive data in the form of graphs or in the form of orientation angles, which can be written to a file or output to display devices.
The library of algorithms that you have collected is truly colossal, but as you correctly understood the problem - it is necessary to make a convenient entry point for using this library.
I am sure that professional programmers write similar algorithms themselves, and it is precisely to popularize your work and these algorithms that ready-made examples of an applied nature need to be made.
Thank you for your keen interest and participation in the discussion!

Mayitzin · 2024-10-01T12:44:52Z

Mayitzin
Oct 1, 2024
Maintainer

Hi!

I added a Jupyter Notebook (https://github.com/Mayitzin/ahrs/blob/master/notebooks/Showcase.ipynb), where I will be showcasing the most common tools of the library, and I will try to make it as accessible as possible. If you think there should be something to further clarify, please let me know so that I can add it.

1 reply

brightproject Oct 2, 2024
Author

I added a Jupyter Notebook

This is great news, thank you for your work.

If you think there should be something to further clarify, please let me know so that I can add it.

I was thinking of adding the ability to transfer raw data from accelerometers and gyroscopes from a flight simulator FlightGear to the algorithm AHRS on a PC and for the algorithm to process these values in Python and output orientation angles to the serial port.
A kind of SITL(Software in the Loop) system.
Now I am processing data from the simulator on a real/physical microcontroller in C++, using the RTIMULib2 algorithm.

RTIMULib2_Null_SimAngles_small.mp4

It would be interesting to run the flight simulator, enable processing in Python using the AHRS library and get orientation angles and configure the filter, add a magnetic field through the WMM(World Magnetic Model), add noise, like real MEMS sensors.
In the code of the AHRS algorithm there is a WMM, as far as I remember, it is possible to use it and obtain a simulation of the magnetic field strength using the rotation matrices of a known magnetic field based on the WMM model and knowing the Euler rotation angles.
It seems to me that such a package would be interesting.
That is, real examples that can be used in practice.
I have a data parser and how to configure the output from the simulator FlightGear, I can send you the files and tell you how to integrate it.
I do not yet have an understanding of how to feed:

Ax, Ay, Az, Gx, Gy, Gz

data to the AHRS code in Python, you understand and sort this out better than me🙂

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Using scripts to process data from the FlightGear simulator #121

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Replies: 4 comments 2 replies

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Select a reply

Using scripts to process data from the FlightGear simulator #121

brightproject Sep 7, 2024

Replies: 4 comments · 2 replies

brightproject Sep 7, 2024 Author

brightproject Sep 8, 2024 Author

Mayitzin Sep 9, 2024 Maintainer

brightproject Sep 9, 2024 Author

Mayitzin Oct 1, 2024 Maintainer

brightproject Oct 2, 2024 Author

brightproject
Sep 7, 2024

Replies: 4 comments 2 replies

brightproject
Sep 7, 2024
Author

brightproject
Sep 8, 2024
Author

Mayitzin
Sep 9, 2024
Maintainer

brightproject Sep 9, 2024
Author

Mayitzin
Oct 1, 2024
Maintainer

brightproject Oct 2, 2024
Author