20 Kalman KK - terrytaylorbonn/auxdrone GitHub Wiki

24.0910

Been making progress with demos (docx's) EVEN WITH KYIV BLACKOUTS!!!, figuring out the kalman math.

Content:

0 Kalman concepts (my viewpoint)
1 My main kalman working docs
2 OTHER KALMAN WORKING DOCS
3 Matlab kalman (From Epic 4 24.0831)
other
- 4 Python
- 5 Navigation (vectornav)

https://www.kalmanfilter.net/default.aspx

https://grpc.io/docs/languages/python/basics/

Kalman Filter Explained With Python Code Robo Code Insights https://www.youtube.com/watch?v=jn8vQSEGmuM&t=575s

0 Kalman concepts (I will redo this section soon...)

Kalman is (I think; I've not talked with any Kalman guru about this) the core python algos that I need to understand to create the final missing autonomous AI drone puzzle piece: The algo that will compute a smooth flight trajectory to a moving target (like a good pilot now does). I think both the Ukrainians/Russians are working on this for kamikadze drones now.

I was (for a short time) working on a drone project (for an American NGO) for a fiber optic drone for the Ukrainian front. The fiber optics would prevent the drone from being jammed. From what I have heard, there are several reasons why this approach is not worth it. I would think the real solution would be to simply have

The copter moves fast (and low) all the time to avoid being in the sites of some "drone-downing gun" for a significant period of time.
When jammed (comms to pilot are cut off), the drone go 100% autonomous. For this it needs a Kalman filter to make a smooth flight pursuit. Something like diagram below (in this situation, the target (gates) dont move, the copter just computes the flight path; only pic I have on hand, would imagine many concepts are same with Kalman).

Note above: Thats really sophisticated. And the UKR/RU military is for sure doing some sophisticated stuff. I am, of course, not trying to "compete" with them, not trying to build something to give the Ukrainians (like that American NGO), but rather to learn the basics, taking inspiration from what they are probably doing.

1 My main kalman working docs in (click here to view in gdrive)

These 2 docs for the dot tracking algo:

#1a visual computer (VC)
#1b NickNair

Hands on great videos

#2a cppmonk (vscode)
#2b rlabbe (notebook)

not done yet, these are more uni type...

#3a hk
#3b toronto
#3c (pdf reference)

#1a 21_33_kalman_motion_prediction_VC_.docx

Great simple demo

tracks a red dot moving in front of 920 camera
kalman creates a blue circle that is the predication of next location
No GIT (i typed in what i think is all his code)
Runs no errors on windows (does nothing; using v4l2)
will try on JNano soon

Using Kalman Filter to Predict Motion

VisualComputer https://www.youtube.com/watch?v=VJ-gB9izieU

#1b 20.1_33b_(36)_kalman_opencv_demos_v02_24.0830

nicknair

https://colab.research.google.com/drive/1j5PNpZqCpuVx7-zxJMMQfHY8Ucl-jqD2#scrollTo=zIFHayI5EdSm

https://github.com/NickNair/Multiple-Object-Tracking-using-Kalman-Filter?tab=readme-ov-file

#2a 20.1_35_kalman_cpp_monk_.docx

My version of The best intro youtube to python kalman (has git). step by step , works.

Tracking example (@4200)

At the end of docx I compare his equations to wiki versions and explain in simple english what it all means.

cppmonk

Understand & Code a Kalman Filter in Python and C++ [Intro] CppMonk [Part 2, Python] https://www.youtube.com/watch?v=m5Bw1m8jJuY&list=PLvKAPIGzFEr8n7WRx8RptZmC1rXeTzYtA&index=3

Play all Understanding & Code a Kalman Filter: Python and C++ CppMonk 8 videos 59,910 views Last updated on Aug 23, 2020 https://www.youtube.com/playlist?list=PLvKAPIGzFEr8n7WRx8RptZmC1rXeTzYtA

cbecker/kalman_python_cpp (suspicously similar to cppmonk :) )

https://github.com/cbecker/kalman_python_cpp

#2b 20.1_32_kalman_rlabbe_.docx

Massive jupyter notebook. All examples work on my pc. For in depth study.

#3a 20.1_34_kalman_hkust_.docx

Great conceptual overview. Finance.

#3b 20.1_31_kalman_toronto_.docx

Coursera course. Can take for free?? not tried it yet. self-driving cars.

2 OTHER KALMAN WORKING DOCS

Visually Explained: Kalman Filters

Visually Explained https://www.youtube.com/watch?v=IFeCIbljreY

Kalman Filter for Beginners

Augmented AI https://www.youtube.com/watch?v=bm3cwEP2nUo

shane ross: Kalman Filter for Beginners, Part 1 - Recursive Filters & MATLAB Examples

Dr. Shane Ross https://www.youtube.com/watch?v=HCd-leV8OkU

State Estimation and Localization for Self-Driving Cars

This course is part of Self-Driving Cars Specialization

https://www.coursera.org/learn/state-estimation-localization-self-driving-cars

Designing Kalman Filters

https://github.com/rlabbe/Kalman-and-Bayesian-Filters-in-Python/blob/master/08-Designing-Kalman-Filters.ipynb

3 Matlab kalman (From Epic 4 24.0831)

https://www.mathworks.com/help/vision/ug/using-kalman-filter-for-object-tracking.html

https://www.mathworks.com/help/vision/tracking-and-motion-estimation.html?s_tid=CRUX_lftnav

4 Python

5 Navigation (vectornav)

https://www.vectornav.com/resources/inertial-navigation-primer