Automatic Label Creation Framework for FMCW Radar Images Using Camera Data

Mèndez, Javier; Schoenfeldt, Stephan; Tang, Xinyi; Valtl, Jakob; Cuéllar, Manuel Pegalajar; Morales, Diego P.

doi:10.1109/access.2021.3087207

Cited by 9 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditionally, radar data is preprocessed in order to extract high level features from the detected targets, such as range and speed. The most popular technique to extract these features is the generation of the RDM using a double Fourier transformation to study the different samples measured by each of the antennas of the radar sensor [3]. This data can be further preprocessed to extract the RAM from each of the detected clusters in the RDM using the beamforming algorithm [6] or MUSIC algorithm [7] among others.…”

Section: State Of the Artmentioning

confidence: 99%

“…The selection of the preprocessing technique depends on the input data type as well as the purpose of the application. In the case of frequency modulated continuous wave (FMCW) radar sensors, the raw data is often preprocessed to extract the Rangle-Doppler Map (RDM) or Rangle-Angle Map (RAM) [3]. These formats are often selected due to their easy feature representation in contrast with the raw radar data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Investigation for the Need of Traditional Data-Preprocessing when Applying Artificial Neural Networks to FMCW-Radar Data

Valtl

Mèndez

Mauro

et al. 2022

2022 29th International Conference on Systems, Signals and Image Processing (IWSSIP)

Self Cite

View full text Add to dashboard Cite

Robust functionality of autonomous driving vehicles relies on their ability to detect obstables and various scenarios on the road. This can be only achieved by applying robust, fast and efficient AI-based signal processing to radar data. In this work we present an empirical investigation on the question, whether one can apply artificial neural networks (ANNs) directly to frequency modulated continuous wave (FMCW) radar raw data. We show that preproceessing is not necessary if one has enough raw data.In our experiment we have data of 153 648 frames collected with a 60 GHz FMCW radar. We compare systematically the options of preprocessing the data using variational autoencoder, applying traditional preprocessing or omit data-preprocessing and apply ANN directly to raw data. We show that the last option results in 28% faster signal processing and highest accuracy. This is a promising result, since it enables edge computing and direct signal processing at the sensor level.

show abstract

Section: State Of the Artmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Investigation for the Need of Traditional Data-Preprocessing when Applying Artificial Neural Networks to FMCW-Radar Data

Valtl

Mèndez

Mauro

et al. 2022

2022 29th International Conference on Systems, Signals and Image Processing (IWSSIP)

Self Cite

View full text Add to dashboard Cite

show abstract

“…The parameter p defines the fractional multiple range. Substituting this into Equation ( 2) results in the simplified form presented as Equation (8). It signifies the sinc shift by K. When p equals zero, S f (n) exhibits zero-crossing sidelobes.…”

Section: Proposed Algorithmmentioning

confidence: 99%

“…The IBDE method employs supervised learning which necessitates acquiring and organizing a sufficient amount of data for each label. Consequently, the issue of labeling has been raised, particularly regarding cost, leading to the investigation of various methods [ 8 , 9 , 10 ]. FMCW radar also encounters limitations owing to the labeling issue.…”

Section: Introductionmentioning

confidence: 99%

FMCW Radar Sensors with Improved Range Precision by Reusing the Neural Network

Cho,

Jung,

Lee

2023

Sensors

View full text Add to dashboard Cite

This paper addresses the challenge of enhancing range precision in radar sensors through supervised learning. However, when the range precision surpasses the range resolution, it leads to a rapid increase in the number of labels, resulting in elevated learning costs. The removal of background noise in indoor environments is also crucial. In response, this study proposes a methodology aiming to increase range precision while mitigating the issue of a growing number of labels in supervised learning. Neural networks learned for a specific section are reused to minimize learning costs and maximize computational efficiency. Formulas and experiments confirmed that identical fractional multiple patterns in the frequency domain can be applied to analyze patterns in other FFT bin positions (representing different target positions). In conclusion, the results suggest that neural networks trained with the same data can be repurposed, enabling efficient hardware implementation.

show abstract