Analysis of the Sustained Performance of Deep Learning Models for Driver Assistance in Mobile SoCs
Resumo
Sistemas Avançados de Assistência ao Condutor (ADAS) aumentam a segurança viária, e a Inteligência Artificial de Borda (Edge AI) viabiliza sua implementação em smartphones sem hardware dedicado. Este estudo avalia o desempenho sustentado do modelo de deep learning SSD-MobileNet v1 para ADAS passivo em dispositivos comerciais, focando em desempenho computacional, eficiência energética e restrições térmicas. Utilizando benchmarks padronizados e uma aplicação Android customizada, analisou-se três níveis de Systems-on-Chip (SoCs): Snapdragon 8 Elite, Helio G99 Ultra e Exynos 1330. Os resultados revelam que dispositivos de entrada e intermediários sofrem fallback compulsório para CPU devido à imaturidade de drivers, enquanto plataformas flagship enfrentam severo throttling térmico sob carga contínua.
Referências
Agência Nacional de Transportes Terrestres (2023). Boletim anual de sinistros de trânsito em rodovias 2023. Technical report, ANTT, Superintendência de Infraestrutura Rodoviária, Brasília.
Badidi, E. (2022). Edge ai and blockchain for smart sustainable cities: Promise and potential. Sustainability, 14(13).
Carvalho, C. H. R. d. (2023). Balanço da 1ª década de ação pela segurança no trânsito no brasil e perspectivas para a 2ª década. Technical report, Instituto de Pesquisa Econômica Aplicada (Ipea), Brasília.
CEDOM - Centro de Documentação e Memória do Mercado Segurador (2025). Acidentes de trânsito no brasil: de 2006 a 2025, muitas mudanças. [link]. Acesso em: 01 jun. 2025.
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., and Adam, H. (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation. In Ferrari, V., Hebert, M., Sminchisescu, C., and Weiss, Y., editors, Computer Vision – ECCV 2018, pages 833–851, Cham. Springer International Publishing.
Elhanashi, A., Dini, P., Saponara, S., and Zheng, Q. (2024). Advancements in tinyml: Applications, limitations, and impact on iot devices. Electronics, 13(17).
Han, S., Pool, J., Tran, J., and Dally, W. J. (2015). Learning both weights and connections for efficient neural networks. In Proceedings of the 29th International Conference on Neural Information Processing Systems - Volume 1, NIPS’15, page 1135–1143, Cambridge, MA, USA. MIT Press.
Howard, A., Sandler, M., Chen, B., Wang, W., Chen, L.-C., Tan, M., Chu, G., Vasudevan, V., Zhu, Y., Pang, R., Adam, H., and Le, Q. (2019). Searching for mobilenetv3. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 1314–1324.
Ignatov, A., Malivenko, G., Timofte, R., Zhang, Z., Wang, Y., Huang, Z., Luo, G., and Yu, G. (2021). Fast and accurate single-image depth estimation on mobile devices, mobile ai 2021 challenge: Report. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 2545–2557.
Ignatov, A., Timofte, R., Chou, W., Wang, K., Wu, M., Hartley, T., and Van Gool, L. (2019). AI Benchmark: All about deep learning on smartphones in 2019. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, pages 3617–3635.
Jacob, B., Kligys, S., Chen, B., Zhu, M., Tang, M., Howard, A., Adam, H., and Kalenichenko, D. (2018). Quantization and training of neural networks for efficient integer-arithmetic-only inference. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2704–2713.
Katare, D., Perino, D., Nurmi, J., Warnier, M., Janssen, M., and Ding, A. Y. (2023). A survey on approximate edge ai for energy efficient autonomous driving services. IEEE Communications Surveys & Tutorials, 25(4):2714–2754.
Li, P., Wang, X., Huang, K., Huang, Y., Li, S., and Iqbal, M. (2022). Multi-model running latency optimization in an edge computing paradigm. Sensors, 22(16).
Lin, W., Adetomi, A., and Arslan, T. (2021). Low-power ultra-small edge ai accelerators for image recognition with convolution neural networks: Analysis and future directions. Electronics, 10(17).
Masello, L., Castignani, G., Sheehan, B., Murphy, F., and McDonnell, K. (2022). On the road safety benefits of advanced driver assistance systems in different driving contexts. Transportation Research Interdisciplinary Perspectives, 15:100670.
Musa, S. S., Zennaro, M., Libsie, M., and Pietrosemoli, E. (2022). Convergence of information-centric networks and edge intelligence for iov: Challenges and future directions. Future Internet, 14(7):192.
Oz, W. (2023). Berlin - germany 4k driving tour 2023 — potsdamer platz, charlottenburg, wilmersdorf, spandau. Vídeo no YouTube. Disponível em: [link]. Acesso em: 28 jun. 2025.
Renault (2024). Sistema de assistência ao condutor (adas) - programa human first. [link]. Acesso em: 23 ago. 2024.
SAE International (2021). Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles. SAE Technical Standard, J3016 202104.
Tan, M. and Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the 36th International Conference on Machine Learning, pages 6105–6114.
Utah, J. (2023). Chicago 4k - night drive - driving downtown. Vídeo no YouTube. Disponível em: [link]. Acesso em: 28 jun. 2025.
Wang, X., Tang, Z., Guo, J., Meng, T., Wang, C., Wang, T., and Jia, W. (2025). Empowering edge intelligence: A comprehensive survey on on-device ai models. ACM Comput. Surv., 57(9).
Weig, F. W. (2016). Advanced driver-assistance systems: Challenges and opportunities ahead. Technical report, McKinsey & Company.
Wulfert, L., Kühnel, J., Krupp, L., Viga, J., Wiede, C., Gembaczka, P., and Grabmaier, A. (2024). Aifes: A next-generation edge ai framework. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(6):4519–4533.
Badidi, E. (2022). Edge ai and blockchain for smart sustainable cities: Promise and potential. Sustainability, 14(13).
Carvalho, C. H. R. d. (2023). Balanço da 1ª década de ação pela segurança no trânsito no brasil e perspectivas para a 2ª década. Technical report, Instituto de Pesquisa Econômica Aplicada (Ipea), Brasília.
CEDOM - Centro de Documentação e Memória do Mercado Segurador (2025). Acidentes de trânsito no brasil: de 2006 a 2025, muitas mudanças. [link]. Acesso em: 01 jun. 2025.
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., and Adam, H. (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation. In Ferrari, V., Hebert, M., Sminchisescu, C., and Weiss, Y., editors, Computer Vision – ECCV 2018, pages 833–851, Cham. Springer International Publishing.
Elhanashi, A., Dini, P., Saponara, S., and Zheng, Q. (2024). Advancements in tinyml: Applications, limitations, and impact on iot devices. Electronics, 13(17).
Han, S., Pool, J., Tran, J., and Dally, W. J. (2015). Learning both weights and connections for efficient neural networks. In Proceedings of the 29th International Conference on Neural Information Processing Systems - Volume 1, NIPS’15, page 1135–1143, Cambridge, MA, USA. MIT Press.
Howard, A., Sandler, M., Chen, B., Wang, W., Chen, L.-C., Tan, M., Chu, G., Vasudevan, V., Zhu, Y., Pang, R., Adam, H., and Le, Q. (2019). Searching for mobilenetv3. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 1314–1324.
Ignatov, A., Malivenko, G., Timofte, R., Zhang, Z., Wang, Y., Huang, Z., Luo, G., and Yu, G. (2021). Fast and accurate single-image depth estimation on mobile devices, mobile ai 2021 challenge: Report. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 2545–2557.
Ignatov, A., Timofte, R., Chou, W., Wang, K., Wu, M., Hartley, T., and Van Gool, L. (2019). AI Benchmark: All about deep learning on smartphones in 2019. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, pages 3617–3635.
Jacob, B., Kligys, S., Chen, B., Zhu, M., Tang, M., Howard, A., Adam, H., and Kalenichenko, D. (2018). Quantization and training of neural networks for efficient integer-arithmetic-only inference. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2704–2713.
Katare, D., Perino, D., Nurmi, J., Warnier, M., Janssen, M., and Ding, A. Y. (2023). A survey on approximate edge ai for energy efficient autonomous driving services. IEEE Communications Surveys & Tutorials, 25(4):2714–2754.
Li, P., Wang, X., Huang, K., Huang, Y., Li, S., and Iqbal, M. (2022). Multi-model running latency optimization in an edge computing paradigm. Sensors, 22(16).
Lin, W., Adetomi, A., and Arslan, T. (2021). Low-power ultra-small edge ai accelerators for image recognition with convolution neural networks: Analysis and future directions. Electronics, 10(17).
Masello, L., Castignani, G., Sheehan, B., Murphy, F., and McDonnell, K. (2022). On the road safety benefits of advanced driver assistance systems in different driving contexts. Transportation Research Interdisciplinary Perspectives, 15:100670.
Musa, S. S., Zennaro, M., Libsie, M., and Pietrosemoli, E. (2022). Convergence of information-centric networks and edge intelligence for iov: Challenges and future directions. Future Internet, 14(7):192.
Oz, W. (2023). Berlin - germany 4k driving tour 2023 — potsdamer platz, charlottenburg, wilmersdorf, spandau. Vídeo no YouTube. Disponível em: [link]. Acesso em: 28 jun. 2025.
Renault (2024). Sistema de assistência ao condutor (adas) - programa human first. [link]. Acesso em: 23 ago. 2024.
SAE International (2021). Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles. SAE Technical Standard, J3016 202104.
Tan, M. and Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the 36th International Conference on Machine Learning, pages 6105–6114.
Utah, J. (2023). Chicago 4k - night drive - driving downtown. Vídeo no YouTube. Disponível em: [link]. Acesso em: 28 jun. 2025.
Wang, X., Tang, Z., Guo, J., Meng, T., Wang, C., Wang, T., and Jia, W. (2025). Empowering edge intelligence: A comprehensive survey on on-device ai models. ACM Comput. Surv., 57(9).
Weig, F. W. (2016). Advanced driver-assistance systems: Challenges and opportunities ahead. Technical report, McKinsey & Company.
Wulfert, L., Kühnel, J., Krupp, L., Viga, J., Wiede, C., Gembaczka, P., and Grabmaier, A. (2024). Aifes: A next-generation edge ai framework. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(6):4519–4533.
Publicado
19/10/2026
Como Citar
FREITAS, Fabrício Daniel; MIRANDA NETO, Arthur de; CORREA, Luiz Henrique Andrade.
Analysis of the Sustained Performance of Deep Learning Models for Driver Assistance in Mobile SoCs. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 1804-1815.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.31373.
