when i add gryY it didint work well after i deploy and use in Ardunio IDE

[Gokhan-Ergul]

when i add gryY it didn't work well after i deploy and use in Ardunio IDE. it works well only acc sensors. if i add gryY sensor, it works in edge impluse very well but not in ardunio ide.

[profvjreddi]

@Mjrovai looks like a good question for you :)

[Mjrovai]

If I understand correctly, you added a Gyroscope but used the code example for an accelerometer. When adding other sensors, you should add the correct sensor libraries at the start of your Arduino sketch. Try the other provided code example with sensor fusion (merger), or please provide us with more details about your question. Thanks.

[Gokhan-Ergul]

If I understand correctly, you added a Gyroscope but used the code example for an accelerometer. When adding other sensors, you should add the correct sensor libraries at the start of your Arduino sketch. Try the other provided code example with sensor fusion (merger), or please provide us with more details about your question. Thanks.

edge_an issue.docx

i added a file that show the issue with details. Thank you.

[Gokhan-Ergul]

Hi, I uploaded a word file, Did you see that, thank you for your helps. Best regards, gokhan ergul Marcelo Rovai ***@***.***>, 11 Tem 2025 Cum, 20:01 tarihinde şunu yazdı:

…

Please provide us with more details about your question. Thanks. — Reply to this email directly, view it on GitHub <#891 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A7PVVEUABFPFUBQHOCUSDNL3H73XZAVCNFSM6AAAAACBKN43HCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNZTGQ2DOMQ> . You are receiving this because you authored the thread.Message ID: ***@***.*** com>

[Mjrovai]

Hi. I just returned to my country and office. Thanks for the file.
When using sensor fusion with different types of sensors, it is crucial to normalize the data. Note that accelerometer data are in the range of a few m/s², but the gyroscope generates angular speed in the hundreds of degrees/s² range. Because of this, RMS values from Acc will be much lower than those from Gyros, which implies that gyroscopic data has a false importance. To enable data normalization, you can use the "Normalize features" option in the Processing Blocks Generate Features tab. This will learn the mean and standard deviation of each output column during the feature generation step, and apply a normalization step during training and inference.

[Gokhan-Ergul]

Hi, thanks for the reply. I understand it but i have a question, why does it workwell on live classification without normalization. Also, as far as I understand, the data is collected in a normalized manner, but when the model is trained, it is presented as an option, not as normalized. Why? Because if it had not been collected in a normalized manner, it should have worked without any problems because the gyroscope values would still show large values in the testing phase as in the traning phase.

…

On Wed, Jul 16, 2025, 17:16 Marcelo Rovai ***@***.***> wrote: Hi. I just returned to my country and office. Thanks for the file. When using sensor fusion with different types of sensors, it is crucial *to normalize the data*. Note that accelerometer data are in the range of a *few m/s²,* but the gyroscope generates angular speed in the *hundreds of degrees/s²* range. Because of this, RMS values from Acc will be much lower than those from Gyros, which implies that gyroscopic data has a false importance. To enable data normalization, you can use the "Normalize features" option in the Processing Blocks Generate Features tab. This will learn the mean and standard deviation of each output column during the feature generation step, and apply a normalization step during training and inference. image.png (view on web) <https://github.com/user-attachments/assets/2dcada6b-a8cf-49c0-9c7c-cf52df7acc78> — Reply to this email directly, view it on GitHub <#891 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A7PVVEQ4UQ76RBPQKHXY3B33IZUFDAVCNFSM6AAAAACBKN43HCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNZXHA3TAMA> . You are receiving this because you authored the thread.Message ID: ***@***.*** com>