Research on the design of multi-functional dual 5G AI edge gateways

doi:10.23919/JSEE.2026.000083

Abstract

Abstract:

With the continuous improvement of 5G network infrastructure, the informatization process in the industrial sector is accelerating at an unprecedented pace. The market demand for gateways integrating 5G networks with multi-access edge computing (MEC) technology is growing increasingly strong. There is also a rising requirement for these gateways to possess capabilities such as high reliability, large bandwidth, fast processing speed, and localized data processing. This paper, in response to the practical management needs of photovoltaic power stations and industrial fields, proposes the design of a 5G artificial intelligence (AI) edge computing intelligent gateway with dual 5G modules. The design is based on the i.MX8M Plus processor and integrates dual 5G module design. It supports multiple communication modes, including 5G, Wi-Fi, Ethernet, and serial ports. The software incorporates edge computing scripts, various communication protocols, and a lightweight AI model, enabling real-time data acquisition and processing at the edge. It supports for several industrial standard protocols ultra-low latency acquisition, and intelligent operation features like remote over-the-air upgrade, secure shell secure operation, and configuration hot updates. The system was tested and validated in a typical photovoltaic application scenario. The results demonstrate that, compared with the traditional single-module Internet of Things gateways, the dual 5G AI edge gateway proposed in this paper reduces end-to-end latency by more than 70%, improves data reliability tenfold, enhances bandwidth by 100%–200%, extends coverage capability by 50%, effectively alleviates the bandwidth pressure on the core network, and enhances the overall reliability and security of the system.

Key words: industrial Internet of Things, gateway, dual 5G, artificial intelligence (AI) edge computing

Fuyan Huang, Gang Xiao, Yongzhong Huang, Xueming Shen, Jiajing Qiu. Research on the design of multi-functional dual 5G AI edge gateways[J]. Journal of Systems Engineering and Electronics, 2026, 37(4): 1161-1170.

Figures/Tables 9

Fig 1

Table 1

Fig 2

Fig 3

Fig 4

Fig 5

Fig 6

Table 2

Fig 7

References 17

1	Liu F, Cui Y, Masouros C, et al Integrated sensing and communications: toward dual-functional wireless networks for 6G and beyond. IEEE Journal on Selected Areas in Communications, 2022, 40 (6): 1728. doi: 10.1109/JSAC.2022.3156632
2	Tang S, Chen B, Iwen H, et al. VECFrame: a vehicular edge computing framework for connected autonomous vehicles[C]// Proc. of the IEEE International Conference on Edge Computing, 2021: 68.
3	Ra E K, Durisi G, Coleri S, et al Wireless 6G connectivity for massive number of devices and critical services. Proceedings of the IEEE, 2025, 113 (9): 826. doi: 10.1109/JPROC.2024.3484529
4	Zhao H W, Zhang Z Q, You J Y, et al Unloading strategy of edge computing in emergency management scenario. Computer Engineering and Design, 2022, 43 (9): 2549.
5	She R, Wu Y, Cui E, et al Energy-efficiency optimization for heterogeneous computing-assisted NOMA-MEC edge AI tasks. Future Generation Computer Systems: the International Journal of Science, 2025, 162, 107458. doi: 10.1016/j.future.2024.07.036
6	Bai Z L, Zhao L, Lu L, et al. Multi-attribute double auction based resource trading in parked vehicle-assisted edge computing[C]// Proc. of the 2nd International Workshop on Radio Frequency (RF) Computing, 2025. DOI: 10.1145/3737906.3767099.
7	Zhang Z Y, Gu K Y, Xu Z J DRL-based task and computational offloading for internet of vehicles in decentralized computing. Journal of Grid Computing, 2024, 22 (1): 3.
8	Ashfaq M, Nur S IoT sensor networks-orchestrating connectivity, efficiency, and intelligence across diverse domains. International Journal of Innovative Research in Computer Science and Technology, 2024, 12, 154. doi: 10.55524/ijircst.2024.12.3.26
9	Tang C J, Tian R, Liu B J, et al. Design and implementation of a secure wireless data acquisition system for substation monitoring[C]// Proc. of the 4th International Conference on Sensors and Information Technology, 2024. DOI: 10.1117/12.3029263.
10	Ranjan R, Villari M, Shen H, et al Software tools and techniques for fog and edge computing. Software-Practice & Experience, 2020, 50, 473. doi: 10.1002/spe.2813
11	Hsieh H, Chen J, Bensilmane A 5G virtualized multiaccess edge computing platform for IoT applications. Journal of Network and Computer Applications, 2018, 115, 94. doi: 10.1016/j.jnca.2018.05.001
12	Moradbeikie A, Pasandi H B, Keshavarz A, et al. Towards secure edge-based LoRa WAN for next generation wireless communications[C]// Proc. of the Iberian Conference on Information Systems and Technologies, 2023. DOI: 10.23919/CISTI58278.2023.
13	Arya A, Pahwa K, Gunjan A butterfly optimization approach for improving the performance of futuristic Internet-of-Things. Evolving Systems, 2023, 15 (3): 1057. doi: 10.1007/s12530-023-09539-4
14	Zuo P F, Lin H M, Deng J B, et al. Serving large language models on huowei CloudMatrix384[EB/OL]. [2025-12-08]. http://arxiv.org/abs/2506.12708.
15	Huang T, You S, Wang F, et al DIST+: knowledge distillation from a stronger adaptive teacher. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2025, 47 (7): 5571. doi: 10.1109/TPAMI.2025.3554235
16	Valentina M, Todor G, Evelina P, et al. Intelligent architecture for real-time personal health monitoring in the context of 3G/4G/5G networks[C]// Proc. of the 29th National Conference with International Participation, 2021. DOI: 10.1109/telecom53156.2021.9659732.
17	Yuan H Y, Chen W, King M, et al. An advanced fanless AIoT smart edge gateway with thickness thermal pad design[C]// Proc. of the Inter Society Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, 2022. DOI: 10.1109/therm54085.2022.9899609.

Parameter type	Parameter description	WEC-SMX8MP
Central processing unit	Clock speed	NXP i.MX 8M Plus/ 1.60 GHz
Central processing unit	Core configuration	Four ARMCortex-A53+ one ARM Cortex-M7
Memory	Type	LPDDR4L 3200MT/s
Memory	Capacity	8 GB
Storage	Type	eMMC 5.1
Storage	Capacity	32 GB
Display	Interfaces	1×HDMI, 2×LVDS, 1×MIPI-DSI
Ethernet	Gigabit/fast Ethernet	4 ports
Input/ output	USB 3.0	2 ports (including 1 boot port)
	UART	4×TTL
	PCIe	1×PCIe 2.0 ×1 lane
	USB2.0	4 ports
Operating system	−	Android, Linux, FreeRT OS
Power supply	Input voltage	9-12V, DC
Environmental specifications	Operating temperature	Standard: −20 ~ 60 °C; Industrial: −40 ~ 85 °C
	Storage temperature	−40 ~ 85 °C
	Operating humidity	10%−95% @40 °C, (non-condensing)
	Storage humidity	0−95% @60 °C, (non-condensing)
Mechanical specifications	Dimensions	82 mm×58 mm

Test dimension	Test metric	RM520（RedCap）	MH5000-31 （5G NR）	Dual-module cooperative test	Test criterion
Basic communication performance	Peak downlink bandwidth	2.1 Gbps	1.28 Gbps	—	Within ±10% of nominal value
	Peak uplink bandwidth	750 Mbps	216 Mbps	—	Within ±10% of nominal value
	Idle end-to-end latency (ping)	28 ms (±3 ms jitter)	15 ms（±2 ms jitter）	—	Industrial grade ≤ 50 ms
	Fully-loaded end-to-end latency (ping)	45 ms (±5 ms jitter)	22 ms（±3 ms jitter）	—	Industrial grade ≤ 80 ms
	24-hour continuous transmission packet loss rate	0.08%	0.02%	—	Industrial grade ≤ 0.1%
	Signal reception sensitivity (RSRP)	−118 dBm	−123 dBm	—	Lower (more negative) indicates better performance
	Reconnection success rate (after network interruption)	99.8% (average reconnection time 1.2 s)	99.95% (average reconnection time 0.8 s)	—	≥99.8%
Core functions of the dual-module system	Primary-backup switchover delay	—	—	85 ms (service uninterrupted)	Industrial grade ≤ 100 ms
	Load balancing traffic distribution deviation	—	—	7% (target 1:1, actual 53:47)	≤10%
	Aggregate downlink bandwidth of dual modules	—	—	3.38 Gbps (interference-free aggregation)	Within ±5% of the sum of single-module bandwidths
	Heterogeneous slice binding accuracy	—	—	100% (uRLLC bound to MH5000, eMBB bound to RM520D)	100%
Long-term stability	72-hour high-load continuous operation	No disconnection/no sudden packet loss increase, logs normal	No disconnection/no sudden packet loss increase, logs normal	No switchover failure, no service interruption, logs normal	No failure/no abnormality
Long-term stability	SIM card insertion/removal reliability	100 insertions/removals, recognition success rate 98.3%	100 insertions/removals, recognition success rate 99.7%	—	≥99.5%

[1]	Ce YAN, Yuanqing XIA, Hongjiu YANG, Yufeng ZHAN. Cloud control for IIoT in a cloud-edge environment [J]. Journal of Systems Engineering and Electronics, 2024, 35(4): 1013-1027.
[2]	Lin ZHU, Junjiang LI, Zijie LIU, Dengyin ZHANG. A multi-resource scheduling scheme of Kubernetes for IIoT [J]. Journal of Systems Engineering and Electronics, 2022, 33(3): 683-692.
[3]	Wang Lina & Gu Xuemai. Efficient enhancing scheme for TCP performance over satellite-based internet [J]. Journal of Systems Engineering and Electronics, 2007, 18(1): 177-182.