Optimization of the automotive millimeter-wave radar data processing using the modified MFNN neural network

Cover Page


Cite item

Full Text

Open Access Open Access
Restricted Access Access granted
Restricted Access Subscription or Fee Access

Abstract

BACKGROUND: Recognition and representation of the road scene is a relevant task in the field of autonomous driving. One of the ways to improve the characteristics of the existing sensory hardware is the use of neural networks for signal processing.

AIM: To conduct an experimental study of the possibilities of using the modified MFNN neural network to increase the resolution of a radar with a small number of channels, to compare it with the classical algorithm based on the fast Fourier transform (FFT), and, in addition, to compare the results with the data obtained from other types of sensors (lidars).

METHODS: The algorithm of the road scene representation, in particular, the detection of pedestrians and cars, is used with a millimeter-range automotive radar and a method for determining data components in relation to the DOA problem, based on the MFNN neural network, modified for the case of representing signals taking complex values in the form of excessive basis coefficients minimizing these coefficients according to the L1 norm. The algorithm based on the fast Fourier transform (FFT) is used for comparative analysis.

RESULTS: As a result of the conducted research, confirmation of the practical feasibility of the developed modifications of the MFNN method was obtained, and the advantage of using a neural network, consisting in increasing the degree of detail of objects, the accuracy of determining their shape and position using a radar with a small number of channels, was demonstrated.

CONCLUSION: The obtained results can be used to develop solutions to improve the efficiency of detecting obstacles on the way of transport, automatic vehicle control, continuous environmental monitoring, and so on, which helps to improve the safety and efficiency of highly automated and autonomous systems.

Full Text

Restricted Access

About the authors

Nikolay V. Panokin

Moscow Polytechnic University

Author for correspondence.
Email: pan.itl.np@gmail.com
ORCID iD: 0000-0003-0380-3733
SPIN-code: 1055-5884

Cand. Sci. (Engineering), Head of the Center for Advanced Development of Autonomous Systems

Russian Federation, Moscow

Ivan A. Kostin

Moscow Polytechnic University

Email: kostin.ivan.a@gmail.com
ORCID iD: 0000-0002-9069-9198
SPIN-code: 6948-1058

Research Scientist of the Center for Advanced Development of Autonomous Systems

Russian Federation, Moscow

Alexander V. Karlovskiy

Moscow Polytechnic University

Email: a.karlovskiy@yandex.ru
ORCID iD: 0000-0001-7660-3375
SPIN-code: 6948-1864

Research Scientist of the Center for Advanced Development of Autonomous Systems

Russian Federation, Moscow

References

  1. Denisov VP, Dubinin DV, Erofeev DV. Using the Maximum-Likelihood Method for the Angular Resolution of the Signal Sources in Multibase Phase Direction Finders. Radiophys Quantum Electronics. 2021;64(7):511–519. doi: 10.1007/s11141-022-10152-9
  2. Krim H, Viberg M. Two decades of array signal processing research-The parametric approach. IEEE Signal Process. Mag. 1996;13:67–94. doi: 10.1109/79.526899
  3. Lee EK, Lee J. Performance analysis of conventional beamforming algorithm for angle-of-arrival estimation under measurement uncertainty. Int. J. Antennas Propag. 2020;2020:1–24. doi: 10.1155/2020/7515139
  4. Capon J. High resolution frequency-wavenumber spectrum analysis. Proc. IEEE. 1969;57:1408–1418. doi: 10.1109/PROC.1969.7278
  5. Cho S, Lee J. Efficient Implementation of the Capon beamforming using the Levenberg–Marquardt scheme for two-dimensional AOA estimation. Prog. Electromagn. Res. 2013;137:19–34. doi: 10.2528/PIER12122711
  6. Schmidt RO. Multiple emitter location and signal parameter estimation. In: Proc. Radc Spectr. Est. Work. Shop. 1979;34:243–258.
  7. Furletov YM, Ivanov AM, Shadrin SS, Toporkov MA. Sound Source Direction of Arrival Estimation for Autonomous Driving Applications. In: 2022 Intelligent Technologies and Electronic Devices in Vehicle and Road Transport Complex (TIRVED), Moscow. 2022:1–5. doi: 10.1109/TIRVED56496.2022.9965523
  8. Cichocki A, Unbehauen R. Neural Networks for Solving Systems of Linear Equations and Related Problems. IEEE Transactions on Circuits and Systems I Fundamental Theory and Applications. 1992;39(2):619–633. doi: 10.1109/81.167018
  9. Zala CA, Barrodale I, Kennedy JS. High- resolution signal and noise field estimation using the L1 (least absolute values) norm. IEEE J. Oceanic Eng. 1987;OE-12:253–264.
  10. Bandler JW, Kellerman W, Madsen K. A nonlinear L1 optimization algorithm for design, modeling, and diagnosis of networks. IEEE Trans. Circuits Syst. 1987;34:174–189.
  11. Abdelmalek NN. Solutions of minimum time problem and minimum fuel problem for discrete linear admissible control systems. Int. J. Syst. Sci. 1978;8:849–859.
  12. Levy S, Walker C, Ulrych TJ, Fullagar PK. A linear programming approach to the estimation of the power spectra of harmonic processes. IEEE Trans. Acoust. Speech. Signal Processing. 1992;30:675–679.
  13. Gribonval R, Morten N. Sparse decompositions in “incoherent” dictionaries. In: Proceedings 2003 International Conference on Image Processing. Barcelona, 2003;1:1–33.
  14. la Cour-Harbo A. Fast Estimation of Optimal Sparseness of Music Signals. In: Proceedings of SPRRA 2006. The Third IASTED International Conference on Signal Processing, Pattern Recognition, and Applications. Innsbruck: ACTA Press, 2006.
  15. Donoho DL, Elad M. Optimally sparse representation in general (nonorthogonal) dictionaries via l1 minimization. In: Proceedings of the National Academy of Sciences of the United States of America. 2003;100(5):2197–2202. doi: 10.1073/pnas.0437847100
  16. Fuchs J-J. On sparse representations in arbitrary redundant bases. IEEE Transactions on Information Theory. 2004;50:1341–1344. doi: 10.1109/TIT.2004.828141
  17. Dantas C, Gribonval R. Stable safe screening and structured dictionaries for faster L1 regularization. IEEE Transactions on Signal Processing. 2019. doi: 10.1109/TSP.2019.2919404
  18. Traonmilin Y, Vaiter S, Gribonval R. Is the 1-norm the best convex sparse regularization? In: iTWIST’18 — international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques, Nov 2018, Marseille, France. Marseille, 2018:1–11.
  19. Bertin N, Daudet L, Emiya V, Gribonval R. Compressive Sensing in Acoustic Imaging. In: Boche H, Calderbank R, Kutyniok G, Vybíral J. (eds) Compressed Sensing and its Applications. Applied and Numerical Harmonic Analysis. Birkhäuser, Cham: Springer, 2015. doi: 10.1007/978-3-319-16042-9_6
  20. Foucart S, Gribonval R. Real versus complex null space properties for sparse vector recovery. Comptes Rendus Mathematique. 2010;348(15–16):863–865. doi: 10.1016/j.crma.2010.07.024
  21. Stanković L, Sejdić E, Stanković S, et al. A Tutorial on Sparse Signal Reconstruction and Its Applications in Signal Processing. Circuits Syst Signal Process. 2019;38:1206–1263. doi: 10.1007/s00034-018-0909-2
  22. Zhishun W, Cheung JY, Xia YS, Chen J. Minimum Fuel Neural Networks and Their Applications to Overcomplete Signal Representations. IEEE Transactions on Circuits and Systems I Fundamental Theory and Applications. 2000;47(8):1146–1159. doi: 10.1109/81.873870
  23. Panokin NV, Averin AV, Kostin IA, et al. Method for Sparse Representation of Complex Data Based on Overcomplete Basis, l1 Norm, and Neural MFNN-like Network. Applied Sciences. 2024;14(5). doi: 10.3390/app14051959

Supplementary files

Supplementary Files
Action
1. JATS XML
Download
2. Fig. 1. The MFNN neural network architecture.

Download (53KB)
Indexing metadata
3. Fig. 2. Functional diagram of the bench.

Download (29KB)
Indexing metadata
4. Fig. 3. An experimental road scene with two closely spaced pedestrians at the same range and with a remote corner reflector: а — frame of the video image; b — the result of the FFT algorithm; c — the result of the excessive basis algorithm.

Download (95KB)
Indexing metadata
5. Fig. 4. An experimental road scene with two pedestrians walking side by side at close ranges and with a remote corner reflector: а — frame of the video image; b — the result of the FFT algorithm; c — the result of the excessive basis algorithm.

Download (97KB)
Indexing metadata
6. Fig. 5. An experimental road scene with the RAV4 car: а — frame of the video image; b — the result of the FFT algorithm; c — the result of the excessive basis algorithm.

Download (94KB)
Indexing metadata
7. Fig. 6. Visualization of a cloud of points formed for the experimental road scene with the RAV4 car: а — the results obtained from the lidar; b — the result of processing the radar data.

Download (134KB)
Indexing metadata

Copyright (c) 2024 Eco-Vector


 СМИ зарегистрировано Федеральной службой по надзору в сфере связи, информационных технологий и массовых коммуникаций (Роскомнадзор).
Регистрационный номер и дата принятия решения о регистрации СМИ: ПИ № ФС 77 - 81900 выдано 05.10.2021.