- Noise filtering - You can use lowpass, bandpass, or any other filtering option
- Clutter removal - You can try Moving Target Indication filters or background subtraction to remove stationary objects
- Range processing - Convert time-domain signal to frequency-domain and analyze range using Fast Fourier Transform, Hamming, or Hann
- Doppler processing - Use FFT along the pulse repetition interval to get Doppler frequency shift
- Normalization - Scale the features to have mean and standard deviation of zero and one respectively
- Data Augmentation - Generate more training data by introducing minor variations
- Dimensionality reduction - Principal Component Analysis
- Training/Validation split
- etc.
How to process raw signal data from FMCW radar RFbeam v-md3 befor using it in classification by depp learning
10 次查看(过去 30 天)
显示 更早的评论
I am currently working on a project involving FMCW radar, specifically the RFbeam v-md3 model. My goal is to process the raw signal data from this radar before using it in a deep learning classification person motion task.
I am looking for advice on the optimal methods and techniques to preprocess raw signal data obtained from the RFbeam v-md3 FMCW radar. What are the essential steps to enhance the quality of the data for subsequent deep learning classification and if any one have already code for this.
I hope someone can give me some instructions and pointers such as how to pre-process (remove noise, do fft/stft) the signal. Sorry for the bad english, Thank You.
I attached the .bin files from the original file in google drive link Output data in google drive
and thank you in advance.
0 个评论
回答(1 个)
Arka
2023-12-26
Some preprocessing steps to refine the raw signal data are mentioned below:
Hope this helps!
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Automotive Radar 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!