Extended Kalman Filter Software Implementation - Sensor Fusion #4 - Phil's Lab #73

Extended Kalman Filter (EKF) implementation and practical considerations. Real-world, real-time implementation and demo on an STM32 microcontroller in C using accelerometer and gyroscope measurements.
Part 4 (final) of sensor fusion video series.

[SUPPORT]

[LINKS]

[TIMESTAMPS]
00:00 Introduction

00:21 Altium Designer Free Trial
00:44 JLCPCB and Design Files

01:06 Pre-Requisites
01:53 'Low-Level' Firmware Overview

07:00 Axis Re-Mapping
08:17 Calibration
09:42 Filtering Raw Measurements

12:12 EKF Algorithm Overview
14:11 EKF Initialisation
17:12 EKF Predict Step

19:26 Matlab/Octave Symbolic Toolbox

21:11 EKF Update Step
22:16 Setting EKF Parameters

23:26 Debug Set-up and Tag-Connect SWD Probe

24:05 Live Demonstration

26:29 Practical Considerations

ID: QIBvbJtYjWuHiTG0uCoK

[dinhtrungche]

Huge respect. I'm an engineering student and you helped me see the clear in a forest of theories. I haven't been able to find such quality information anywhere, thank you so much.

[canyonrider]

Great video series! My master thesis was all about an error state space Kalman filter for the use on a multicopter, estimating the accelerometer and gyroscope errors and bias using different measurements. It’s really interesting to see that running on real hardware, rather than simulation. If you’re planning on doing a video on error state space, I‘d recommend using Simulink for the algorithms and Embedded Coder with the Hardware Support Packages for Arm Cortex-M.

[isaacclark]

One of my favorite topics! Well worth the wait! I've not found very much practical EKF discussion for free. Lots of theoretical stuff...

[kalaghori]

You've actually got so quality stuff on your channel. Keep that up man!

[fredo]

I waited a long time for this one since the last part of the series! Thank you!

[mohammadmalaktammo]

Thank you very much! I am doing now my thesis on this topic and I have learned from your videos more than from the 3 scientific papers I read😊

[chinoramas]

Thank you! I'm waiting for this part 4 video. Currently, I have a thesis regarding PDR. These videos really help!!! Thank you!

[nhlakaniphombatha]

THIS GUY IS THE BEST ..WE ALL LOVE YOU PHIL ❤❤🙌🙌🙌

[steev]

Thanks for the video Phil!

Currently I use a complementary filter for my flight control software and it works quite well. But one day I may go down this rabbit hole...

[gretarmark]

Amazing and very inspiring video. I studied this theory in my master's program where we only used Matlab and Simulink. It's nice to see hands-on example like this! It would be interesting to compare the Extended Kalman Filter to the Unscentend Kalman Filter. I'm not sure which one is more robust.

[vycka]

Finally an end to this topic, been waiting for so long

[joegibbs]

Fantastic video! My background is state estimation algorithms but I've always avoided deploying them to actual hardware, this is a fab tutorial on making the leap from Matlab/Simulink to C!

[Elliot_]

So in non sensor fusion applications of EKF, the prediction step is purely based on theoretical equations/models, and then the update step incorporates real measured data. In this sensor fusion example, why is the real gyroscope measured data not used in the update step? And why is it relevant to the theoretical prediction step? Doesn't this somewhat defeat the purpose of a Kalman Filter? Thankyou.

[arielvieiralimaserafim]

Can you release the source code ?, I did not find it in your GitHub.

[joaquinmarianopineiro]

Hello, why there is a changing in sign for x and y coordinates in acceleration? The direction of x+ or y- is aligned with mpu

[isaacclark]

Should SPI be used instead of I2C for mission-critical systems? If not, what would you recommend?

[RupertBruce]

Nice trick using Octave to derive the jacobian

[oguzkaan]

Thank you awsome topic and you made look it easy. But when I was watching and writing code at the same time it could be annoying. Could we reach the code inside the update function ??

[AustynLoehr]

Could you provide some background on the decision to use gyro readings in the predict step and not just include both gyro and accel data in the update step? I assume this is because it is difficult to come up with a better state transition function without knowing more about the system which is IMU is sensing.

[shreyasacharya]

Great lecture! The following Github link doesn't take me to the source code, or did I make any mistakes?