Smart cycling warning systems are crucial for ensuring cycling safety. However, traditional cycling equipment has several limitations, including large blind spots, one-dimensional warning mechanisms and a lack of individual status monitoring. To address these issues, this paper proposes a smart cycling glasses control system based on dual-core heterogeneous processing. During daily cycling, riders often fail to observe rear-approaching vehicles in time, which leads to frequent sudden collisions. Existing warning devices rely solely on radar to provide basic alerts, resulting in limited effectiveness, and they cannot monitor the rider's physical condition. Furthermore, in the event of a collision, these devices are unable to automatically call for help, which can delay rescue operations. The smart cycling glasses system, which is based on dual-core heterogeneous processing and the lightweight YOLOv5 algorithm, has significant advantages in terms of improving warning capabilities and reducing accident casualties. Therefore, a smart cycling warning glasses system has been designed and implemented to solve these problems.

Yang Chen, Wen Xuan, Nima Puchi, Fang Yangbin, Xu Shuping. Research on Intelligent Cycling Warning Glasses Based on Dual-core Heterogeneous and Lightweight YOLOv5. Academic Journal of Engineering and Technology Science (2026), Vol. 9, Issue 3: 118-126. https://doi.org/10.25236/AJETS.2026.090316.

[1] Xia M, Zhou Z, Wu X, et al. Research on machine learning-based wearable dynamic cycling posture monitoring system[J]. International Journal of Clothing Science and Technology, 2026, 38(2): 277-291. DOI: 10.1108/IJCST-01-2025-0006.

[2] Cardenas C R, Amor D, Gutiérrez Á. Edge AI in Practice: A Survey and Deployment Framework for Neural Networks on Embedded Systems[J]. Electronics, 2025, 14(24): 4877. DOI: 10.3390/ELECTRONICS14244877.

[3] Salah H B O, Messaoud S, Hajjaji A M, et al. Post-training quantization for efficient FPGA-based neural network acceleration[J]. Integration, 2025, 105 102508. DOI: 10.1016/j.VLSI.2025.102508.

[4] Chen Zhaoxi, Zheng Peiyu, Chen Jielin, et al. Research on image recognition system based on YOLOv5-Lite and RK3566[J]. Science and Technology Innovation and Application, 2025, 15(22): 6-10+15. DOI: 10.19981/j.CN23-1581/G3.2025.22.002.

[5] Tao Ran, Wang Weizhen. Research progress of motorcycle helmets based on bibliometrics[J]. Progress in Textile Science and Technology, 2025, 47(07): 17-21+28. DOI: 10.19507/j.cnki.1673-0356.2025.07.012.

[6] He X, Zhang Y, Zhan Q. AIN-YOLO: A lightweight YOLO network with attention-based InceptionNext and knowledge distillation for underwater object detection[J]. Advanced Engineering Informatics, 2025, 66: 103504. DOI: 10.1016/j.AEI.2025.103504.

[7] Yang Guang. Research on the design of wearable devices for cycling based on AR technology[D]. North China University of Technology, 2025. DOI: 10.26926/d.cnki.gbfgu.2025.000821.

[8] Huang J, He W, Luo R, et al. Lightweight EdgeYOLO-CRE network for Jaboticaba tree recognition based on ascend AI processor[J]. Smart Agricultural Technology, 2025, 11: 100991. DOI: 10.1016/j.ATECH.2025.100991.

[9] Lin Haifeng, Liu Dapeng, Cai Hui. Improved YOLOv5s for vehicle detection from traffic camera perspective[J]. Computer and Digital Engineering, 2025, 53(05): 1327-1332+1398.

[10] Qian W, Zhu Z, Zhu C, et al. FPGA-based accelerator for YOLOv5 object detection with optimized computation and data access for edge deployment[J]. Parallel Computing, 2025, 124: 103138. DOI: 10.1016/j.PARCO.2025.103138.

[11] Wang Y, Zhang J, Zhou J. Urban traffic tiny object detection via attention and multi-scale feature driven in UAV-vision[J]. Scientific Reports, 2024, 14(1): 20614. DOI: 10.1038/S41598-024-71074-2.

[12] Feng Peiqi, Zhu Yuancheng, Xie Chuang, et al. A helmet capable of monitoring wearing status[J]. Science and Technology & Innovation, 2024, (09): 62-64. DOI: 10.15913/j.cnki.kjycx.2024.09.017.

[13] Yu Bosong, Lu Meiling. Design of automatic weather station test system based on Bluetooth and WeChat mini program[J]. Electronic Technology, 2024, 53(03): 312-313.

[14] Mario B, Baran M U, T. K G, et al. Using body sensors for evaluating the impact of smart cycling technologies on cycling experiences: a systematic literature review and conceptual framework[J]. European Transport Research Review, 2024, 16(1): DOI: 10.1186/S12544-024-00635-3.

[15] Luo Shijian, Gong Hebo, Lin Wei. Research status and progress of intelligent product interaction design[J]. Journal of Mechanical Engineering, 2023, 59(11): 1-15.

[16] Yu Ze. Research on helmet-wearing detection technology for cyclists based on deep learning[D]. Zhejiang Gongshang University, 2022. DOI: 10.27462/d.cnki.ghzhc.2022.000637.

[17] Liu Ting. Key points of PCB design for smart locks[J]. Electronic Production, 2022, 30(11): 88-90. DOI: 10.16589/j.cnki.cn11-3571/tn.2022.11.026.

[18] Geng Anyu. Optimization of lightweight deep learning models and its application in object detection[D]. Hunan University, 2021. DOI: 10.27135/d.cnki.ghudu.2021.002876.

[19] Dai Leiyan, Feng Jie, Dong Hui, et al. Optimized object detection algorithm based on deep learning for embedded devices[J]. Computer Systems & Applications, 2019, 28(04): 163-169. DOI: 10.15888/j.cnki.csa.006866.

[20] Christie N, Talbot R, Reed S, et al. 296 Cyclist deaths in London: implications from a safe systems perspective[J]. Injury Prevention, 2016, 22(S2): A108. DOI: 10.1136/injuryprev-2016-042156.296.