G.I. Biryukov – Student, Bauman Moscow State Technical University

E-mail: biryukovgi@student.bmstu.ru

V.P. Zhalnin – Ph.D. (Eng.), Bauman Moscow State Technical University

E-mail: zhalninvp@bmstu.ru

D.V. Laptev – Master Student, Bauman Moscow State Technical University

E-mail: dmytry2010@student.bmstu.ru

P.O. Repnikov – Student, Bauman Moscow State Technical University

E-mail: rtfiof@gmail.com

The use of convolutional neural networks (CNN) in the tasks of pattern recognition, noise removal, and increasing the resolution of images allows achieving better results. At the same time, there are a number of unresolved problems – a rather complicated development and debugging of the algorithm, a lack of finished libraries, etc.

One of the main problems of convolutional neural networks is their speed. This requirement is key to many neural network applications that require an almost instant and error free result. One way to increase this criterion is through specialized hardware such as programmable logic integrated circuits (PLIC).

The purpose of the neural network considered by the authors is to process in real time a video stream for recognition of road markings. For such systems, one of the key criteria is speed. They are a key part of advanced driving assistance systems (ADAS).

The considered method for constructing a neural network is implemented in the MATLAB environment, which allows you to simulate various models of hardware. The authors consider the construction of a neural network and its subsequent modeling for PLIC Artix-7 and PC. Comparing the results obtained in various models, the authors conclude the preferred hardware for the neural network.

In this paper is proposed an approach to the development and deployment of a neural network, as well as the use of programmable logic integrated circuits as a platform for its work, which provides faster recognition performance and the possibility of implementation in ASIC format. As an example, the problem of recognizing road markings is considered, which allowed quantitative estimates of the result.

Biryukov G.I., Zhalnin V.P., Laptev D.V., Repnikov P.O. Features of implementation of convolutional neural networks on programmable logic integrated circuit Artix-7. Neurocomputers. 2020. V. 22. № 3. P. 26–35. DOI: 10.18127/j19998554-202003-03

1. Schmidhuber J. Deep learning in neural networks: An overview. Neural Networks. 2015. № 61. P. 85–117.
2. Deng L., Hinton G., Kingsbury B. New types of deep neural network learning for speech recognition and related applications: an overview. IEEE International Conference on Acoustics. Speech and Signal Processing. 2013. P. 8599-8603. DOI: 10.1109/CASSP.2013.6639344
3. Qiu J., Song S., Wang Y., Yang H., Wang J., Yao S., Guo K., Li B., Zhou E., Yu J., Tang T., Xu N. Going deeper with embedded FPGA platform for convolutional neural network. ACM/SIGDA International Symposium on Field-Programmable Gate Arrays. 2016. P. 26-35. DOI: 10.1145/2847263.2847265
4. Shakhnov V.A., Vlasov A.I., Polyakov Yu.A., Kuznetsov A.S. Neyrokompyutery: arkhitektura i skhemotekhnika. M.: Mashinostroyeniye. 2000. 64 s. (in Russian).
5. Satzoda R.K., Trivedi M.M. Vision based Lane Analysis: Exploration of Issues and Approaches for Embedded Realization. IEEE Conference on Computer Vision and Pattern Recognition. 2013. DOI: 10.1109/CVPRW.2013.91
6. Zhang S., Li P., Sun G., Guan Y., Xiao B., Cong J. Optimizing FPGA-based accelerator design for deep convolutional neural networks. ACM/SIGDA International Symposium on Field-Programmable Gate Arrays. 2015. P. 161-170. DOI: 10.1145/2684746.2689060
7. Suda N., Chandra V., Dasika G., Mohanty A., Ma Y., Vrudhula S., Seo J., Cao Y. Throughput-optimized openCL-based FPGA accelerator for large-scale convolutional neural network. ACM/SIGDA International Symposium on Field Programmable Gate. 2016. P. 16-25. DOI: 10.1145/2847263.2847276
8. Duong T.T., Pham C.C., Tran T.H.-P., Nguyen T.P., Jeon J.W. Near real-time ego-lane detection in highway and urban streets. IEEE International Conference on Consumer Electronics-Asia (ICCE-Asia). 2016. DOI:10.1109/ICCE-Asia.2016.7804748
9. Sahin S., Becerikli Y., Yazici S. Neural Network Implementation in Hardware Using FPGAs. In: King I., Wang J., Chan L.W., Wang D. (Eds) Neural information processing. ICONIP. Lecture notes in Computer Science. Springer. 2006. V. 4234. P. 1105-1112.
10. Vlasov A.I. Apparatnaya realizatsiya neyrovychislitelnykh upravlyayushchikh system. Pribory i sistemy. Upravleniye. kontrol. diagnostika. 1999. № 2. S. 61-65. (in Russian).
11. Yudin A.V., Shakhnov V.A., Usov K.A., Salmina M.A., Vlasov A.I. Design methods of teaching the development of internet of things components with considering predictive maintenance on the basis of mechatronic devices. International Journal of Applied Engineering Research. 2017. V. 12. № 20. S. 9390-9396.
12. Kotelnitskiy A.V., Vlasov A.I. Primeneniye povedencheskikh modeley pri proyektirovanii sistem na kristalle. Inzhenernyy vestnik. 2012. № 9. S. 10. (in Russian).
13. Vlasov A.I. Neyrosetevaya realizatsiya mikroprotsessornykh sistem aktivnoy akusto-i vibrozashchity. Neyrokompyutery: razrabotka. primeneniye. 2000. № 1. S. 40-44. (in Russian).
14. Vlasov A.I., Volodin E.A., Sementsov S.G., Shakhnov V.A. Elektronnyye sistemy aktivnogo upravleniya volnovymi polyami: istoriya i tendentsii razvitiya. Zarubezhnaya radioelektronika. Uspekhi sovremennoy radioelektroniki. 2002. № 4. S. 3-23.  (in Russian).
15. Myslovskiy E.V., Vlasov A.I., Kuznetsov A.S. Tsifrovyye signalnyye protsessory s fiksirovannoy tochkoy semeystva ADSP21XX. M.: Izd-vo MGTU im.N.E.Baumana. 2003. T. 2. 75 s. (in Russian).
16. Myslovskiy E.V., Vlasov A.I., Menshov K.A. Tsifrovyye signalnyye protsessory s plavayushchey tochkoy semeystva ADSP2106X. M.: Izd-vo MGTU im. N.E.Baumana. 2003. T. 3. 75 s. (in Russian).
17. Averianikhin A.E., Vlasov A.I., Evdokimova E.V. Ierarkhicheskaya piramidalnaya subdiskretizatsiya v glubokikh svertochnykh setyakh dlya raspoznavaniya vizualnykh obrazov. Neyrokompyutery: razrabotka. primeneniye. 2020. (v pechati). (in Russian).
18. Viryasova A.Yu., Vlasov A.I., Gladkikh A.A. Neyrosetevyye metody defektoskopii integralnykh struktur. Neyrokompyutery: razrabotka. primeneniye. 2019. T. 21. № 2. S. 54-67. DOI: 10.18127/j19998554-201902-06 (in Russian).
19. Vlasov A.I., Papulin S.Yu. Analiz dannykh s ispolzovaniyem gistogrammnoy modeli kombinatsii priznakov. Neyrokompyutery: razrabotka. primeneniye. 2019. T. 21. № 5. S. 18-27. DOI: 10.18127/j19998554-201905-02 (in Russian).
20. Lutsiv V.R. Svertochnyye iskusstvennyye neyronnyye seti glubokogo obucheniya. Opticheskiy zhurnal. 2015. T. 82. № 8. S. 11-23. (in Russian).
21. Erokhin D.Yu., Ershov M.D. Sovremennyye svertochnyye neyronnyye seti dlya obnaruzheniya i raspoznavaniya obyektov. Tsifrovaya obrabotka signalov. 2018. № 3. S. 64-69. (in Russian).
22. Gorelov D.V., Vlasov A.I. Sovremennyye sistemy raspoznavaniya registratsionnykh znakov transportnykh sredstv. Trudy Vosmoy Vseros. konf. Molodykh uchenykh i spetsialistov «Budushcheye mashinostroyeniya Rossii». 2015. S. 389-394. (in Russian).
23. Korshunova K.P. Svertochnyye nechetkiye neyronnyye seti dlya resheniya klassifikatsionnykh zadach. Neyrokompyutery: razrabotka. primeneniye. 2017. № 3. S. 44-51. (in Russian).
24. Akhmetzyanov K.R., Yuzhakov A.A. Uvelicheniye tochnosti svertochnoy neyronnoy seti za schet vozrastaniya kolichestva dannykh. Neyrokompyutery: razrabotka. primeneniye. 2018. № 7. S. 14-19. DOI: 10/18127/j19998554-201807-03 (in Russian).