A Novel Wi-Fi Mesh Network Framework for Efficient Mobile Data

Transmission

Yu-Jie Ou, Yan-Ming Chen and Chun-Chao Yeh

Department of Computer Science and Engineering, National Taiwan Ocean University,

Keelung 202301, Taiwan

Keywords: Wi-Fi Mesh Networks, Mobile Data Transmission, QUIC.

Abstract: The current trend involves the use of robots and various IoT devices to assist in production and development.

In the future, with the integration of AI technology, related applications will become even more widespread.

Within this context, whether it is command transmission, equipment status reporting, or equipment condition

monitoring, a stable network environment is essential. Presently, the technology can be mainly categorized

into wired networks, wireless networks (Wi-Fi), and mobile networks. While wired networks are fast and

stable, they are constrained by physical lines and cannot be used in environments requiring interlaced

movement. Wireless networks face disconnection issues when crossing different APs. While mobile networks,

particularly those using 5G, offer excellent real-time performance, the establishment of private networks is

expensive and the signal sources are relatively singular. This study proposes the use of Wi-Fi technology

combined with the UDP (User Datagram Protocol) and QUIC (Quick UDP Internet Connections) protocols

to create a new type of multipath network system. This system aims to reduce the adverse effects of physical

environmental changes and wireless access point (AP) device transitions, achieving a low-cost, high-speed,

highly mobile, and secure network environment in specific settings. The proposed scheme leverages the

encryption and security features of the QUIC protocol to protect data privacy while supporting the needs of

high-mobility applications.

1 INTRODUCTION

In contemporary industrial production and warehouse

management, efficient communication systems are

crucial for maintaining operational efficiency.

Particularly in large warehouses, manufacturing

environments, and other locations that require high

mobility and flexibility, the network environment

needs to support high-speed data transmission while

being able to cope with the impacts of environmental

interference and equipment transitions. This study

proposes the use of Wi-Fi technology combined with

the UDP (User Datagram Protocol) and QUIC (Quick

UDP Internet Connections) protocols to create a new

type of multipath network system. This system aims

to reduce the adverse effects of physical

environmental changes and wireless access point

(AP) device transitions, achieving a low-cost, high-

speed, highly mobile, and secure network

environment in specific settings.

The main research motivation of this study is to

address the wireless network communication needs of

mobile autonomous devices in future unmanned

factories (such as unmanned warehouses) by

designing a Wi-Fi mobile network data transmission

system with data privacy protection and high

reliability using the QUIC network protocol. Based

on data security (for example, ensuring that data is

stored locally rather than uploaded to third-party ISPs

for storage or processing), we believe that

establishing a dedicated network system is necessary.

Table 1 lists the possible network solutions

currently available. After comprehensive evaluation,

it can be found that WiFi is a good choice. The

popularity and technological maturity of Wi-Fi

networks make them an ideal basis for implementing

such network systems. However, traditional Wi-Fi

systems often perform poorly in the face of

environmental interference and switching between

APs. In addition, with the improvement of network

security requirements and the increase of data

transmission volume, existing Wi-Fi systems face

many challenges in ensuring data integrity and

communication efficiency. Therefore, this study

450

Ou, Y.-J., Chen, Y.-M. and Yeh, C.-C.

A Novel Wi-Fi Mesh Network Framework for Efﬁcient Mobile Data Transmission.

DOI: 10.5220/0013480200003944

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Internet of Things, Big Data and Security (IoTBDS 2025), pages 450-455

ISBN: 978-989-758-750-4; ISSN: 2184-4976

explores the use of UDP's low latency characteristics

and QUIC's encryption and security features to

optimize the performance of Wi-Fi networks. The

main objectives of the study include:

 Enhance reliability of Wi-Fi transmission:

Explore technical solutions to mitigate the

effects of packet loss due to error-prone

wireless transmission.

 Improve AP handoff performance: Develop

seamless AP handoff technology to improve

network stability and reliability in dynamic

environments.

 Achieve high transmission speed and low cost:

By combining efficient protocols and cost-

effective wireless technology, an affordable

and efficient communication solution is

provided.

 Enhance security and mobility: Leverage the

encryption and security features of the QUIC

protocol to protect data transmission from

threats while supporting the needs of high-

mobility applications.

We hope to develop an innovative network

architecture that is applicable to a variety of

commercial and industrial scenarios, which will not

only improve the efficiency of on-site operations, but

also enhance the security and reliability of data

transmission. These achievements will help promote

the development of wireless communication

technology in the future, especially in the growing

fields of smart manufacturing and smart warehouses.

Table 1: Comparisons between different wireless

technologies.

Wi-

LoRa

HaLow BT 5G

Bandwidth

++ - + - ++

Easy to use + - - + +

maintainability + + + + -

robustness - - - - +

Low cost ++ ++ + ++ -

Coverage + ++ ++ - ++

2 PROPOSED SYSTEM

FRAMEWORK: DWM

2.1 DWM System Overview

The main motivation for this research topic is to meet

the wireless network communication needs of mobile

autonomous devices in future unmanned factories

(such as unmanned warehouses). From Table 1, we

can find that Wi-Fi is a relatively good choice. In

order to provide Wi-Fi coverage and improve the

reliability of data transmission, we propose a DWM

(Dual-connected Wi-Fi Mesh) network architecture

(Figure 1). In the proposed DWM network system

architecture, each autonomous mobile device such as

an AGV (Automated Guided Vehicle) or an AMR

(Autonomous Mobile Robots) is equipped with two

Wi-Fi NICs, which are connected to two different

SSID APs (SSID-A and SSID-B). In the AP layout of

the mesh network, we use a grid structure, with an AP

placed at each grid point.

The SSIDs of adjacent APs are set to SSID-A and

SSID-B as in the example shown in Figure 1. Under

this architecture, we allow the mobile clients

(AGV/AMR) move in the network with maintaining

the possibility of connecting to two APs in most

cases. Each mobile client is given an IP address (for

Tun interface), and its IP gateway is designated to the

IP address of the server gateway. At the same time,

two IP addresses are assigned to the two underlying

Wi-Fi wireless network interfaces. Basically, these

three interfaces (TUN x1 + Wi-Fi x2) belong to

different IP subnets.

Figure 1: A 3x3 DWM network architecture.

2.2 DWM Software Architecture

Figure 2 shows the software architecture of DWM.

The communication requirements of AGV/AMR in

unmanned factories are mainly the communication

between each AGV/AMR and the remote central

control server gateway (communication between

AGV/AMR needs to go through the server gateway).

Therefore, we mainly focus on the communication

between individual AGV/AMR and server gateway.

The idea of the architecture is basically borrowed

from the IPsec tunnel architecture. From the

perspective of the AGV/AMR Client App

programmer, it can be regarded as a virtual link

directly connecting the AGV/AMR and the remote

A Novel Wi-Fi Mesh Network Framework for Efﬁcient Mobile Data Transmission

451

server gateway without having to care about the

actual underlying network architecture. Therefore,

the burden on upper-level application developers can

be greatly reduced. They only need to focus on the

messages that need to be reported to the server

gateway or how to process the messages sent by the

server gateway.

In addition, because we use the QUIC API, all

data will be encrypted by QUIC, thus ensuring the

risk of information leakage during data transmission.

At the same time, QUIC provides a TCP-like network

packet resending mechanism, thus ensuring that

transient packet loss would not impair the integrity of

the data. The underlying software architecture

basically adopts the proposed Implicit Multi-Path

QUIC (iMP-QUIC) software architecture introduced

in the next subsection.

Figure 2: DWM software architecture.

2.3 Implicit Multi-Path QUIC

The key idea of iMP-QUIC (Implicit Multi-Path

QUIC) is to separate QUIC from the underlying

multi-path transmission network, and to separate the

client/server program data communication (data

plane) from the underlying multi-path packet

manipulation (control plane). Through the underlying

TUN/TAP virtual network interface, iMP-QUIC

provides a communication tunnel between Client-

Server programs. The application program (data

plane) only needs to focus on application data

exchange/handling between the client and server. The

underlying multi-path packet manipulation (control

plane) is handled by the Tun/Tap process daemon in

the iMP-QUIC architecture. Since the Tun/Tap

process daemon program is a user process (not a

kernel module), it can be easily developed by the

application program developer. The QUIC lib in iMP-

QUIC can be any third-party QUIC protocol stack

without special support for QUIC-MP (multipath

QUIC). From the QUIC perspective, its underlying

layer is a single network interface (TUN/TAP virtual

interface), so it is a standard single network interface

and does not need to run the QUIC-MP function.

As shown in Figure 3, when the upper layer

application data is sent out through the QUIC socket

interface, QUIC encapsulates the data into a QUIC

packet and sends it out. The QUIC packet is routed

through the kernel network protocol stack and sent to

the Tun/Tap virtual interface. The Tun/Tap process

daemon receives the data from the upper layer. When

QUIC packets are sent, they are processed according

to the daemon program settings and then handed over

to the multi-link connection middleware below for

processing. The multi-link connection middleware

provides a (UDP) transport service between

client/server hosts, with one UDP connection for each

path. When data is sent to the other party through the

multi-link connection middleware, the Tun/Tap

process daemon collects the packets from different

UDP connections, processes the packets according to

the settings of the daemon program, and then sends

the processed packets (that is, the packets sent by the

client to the client) to the client. The QUIC packet

sent out by the client is written into the TUN/TAP

virtual interface, and the QUIC packet sent out by the

client is sent to the QUIC socket buffer on the server

side through the system routing, and then handed over

to the server program for processing. And, the packet

transmission process from the server side to the client

side is similar to the above procedure.

Figure 3: Packet transmission flow of iMP-QUIC.

IoTBDS 2025 - 10th International Conference on Internet of Things, Big Data and Security

452

3 EXPERIMENTS

3.1 Experiment Settings

We developed a small prototype system. Due to

limited experimental space, we were unable to

conduct large-scale complete system experiments.

We set up three APs along the corridor in our

department. The positions of the three APs are set

according to their signal coverage areas, to make sure

the signal coverage areas of two immediately adjacent

APs are overlap with each other, as shown in Figure

4. The yellow circles, in Figure 4, refer to coverage

are of the first and third APs with the same SSID

(SSID-A); the red circle refers to the coverage of the

second AP with another SSID (SSID-B). In addition,

we made an Arduino robot car to simulate

AGV/AMR. A laptop was placed on the robot car,

which executed the software for AGV/AMR to

communicate with the server gateway, as shown in

Figure 5. This laptop is equipped with two Wi-Fi

network cards. The equipment for other experiments

is shown in Table 2.

Figure 4: A small scale field try on the corridor.

Figure 5: The Arduino robot car. (a) the control units, (b)

the carrying platform on top of the control units. (c) the NB

mounted on the carrying flatworm.

Table 2: HW/SW specifications of the experiment

components.

Devices Spec.

Server gateway

(PC)

CPU: CORE i7 10

OS: Ubuntu 20.04.6

Client

(Notebook):

CPU: CORE i7 11

OS: Ubuntu 20.04.6

Wi-Fi adaptor 1&2 ASUS USB-AC51 Wireless-AC600

Wi-Fi AP x3 ASUS RT-AX3000

3.2 Performance Evaluations

Since QUIC is a reliable transmission protocol similar

to TCP, we use the existing QUIC package. In our

implementation software, we cannot observe the

actual QUIC packet loss situation directly (as the

underlying QUIC packet is managed by the QUIC

packet itself), so we use an indirect method to observe

the UDP connection in the underlying multi-link

connection middleware. At the same time, for our

system design, we evaluated different system design

options, especially the following points:

 Single link vs. dual link;

 Using the same SSID in dual link (AAA, all

three APs are set to SSID_A) or alternative

SSID (ABA, three APs are set to SSID_A/B/A

in sequence);

 The impact of AGV/AMR moving speed on

transmission performance;

 The impact of client transmission behaviour

(transmission frequency and packet size) on

transmission performance.

Since the client has two Wi-Fi interfaces, in order

to record the client connection to the AP, we record

the AP RSSI strength in each packet sent by the client.

Experimental default parameters: client sends 1

packet per second; vehicle moving speed: Fast =

0.984 m/s (3.5 km/h), Slow = 0.328 m/s (1.18 km/h);

data packet size It is 47 bytes.

3.3 Experimental Results

In this subsection we present the experimental results.

3.3.1 Constant Data Sending Rate (3pps

Sending Rate)

First, we evaluate packet transmission efficiency

regarding packet receiving rate and loss at receiving

site (the server gateway), in which the mobile client

transmits data at a fixed rate of 3pps (3 packets per

second) to simulate sensor data transmission scenario.

A Novel Wi-Fi Mesh Network Framework for Efﬁcient Mobile Data Transmission

453

The data size is 2KB (UDP or QUIC data payload

size). The experimental results are shown in Table 3.

In Table 3, QUIC-ABA/AAA represents the

proposed iMP-QUIC scheme is applied and the

SSIDs of the three Wi-Fi APs are set to the same

(SSID-A for AAA) or interleaving (SSID-A/SSID-

B/SSID-A for ABA); The UDP-ABA represents a

similar scheme as iMP-QUIC, in which we replace

QUIC with UDP in the upper layer. Since QUIC

packets are automatically resent if lost, the value of

pkt_loss_rate is 0. From Table 3, we observe that

when the mobile client moves at fast speed, the QUIC

ABA/AAA suffers more packet lost and

retransmission and thus a lower received packet rate

is experienced in receiver side. This is mainly because

the underlying Wi-Fi connection is interrupted when

switching between APs. When the mobile client

moves at slower speed, the impacts of the connection

interruption is alleviated, and both QUIC-ABA and

QUIC-AAA can achieve receiving packet rate of

close to 3pps, the data generation rate at the sender

side.

Table 3: packet receive rate and loss rate (sender rate=3

pps).

speed QUIC -

ABA

QUIC-

AAA

UDP -

ABA

PPS fas

2.57 2.08 3.02

loss 0 0 0.104

PPS slow 2.94 2.98 2.95

loss 0 0 0.52

3.3.2 Unbounded Data Sending Rate

We, at the same time, simulated the situation of large

file transmission. We had the mobile client transmit

the data continuously without any break between two

consecutive data transmissions. The size of each

transmission is the same at 2KB (UDP or QUIC data

payload size). The experimental results are shown in

Table 4.

We analyse further the packet received at different

layer of the proposed iMP-QUIC protocol (Figure 3):

upper layer (QUIC for QUIC-ABA/-AAA, and UDP

for UDP-ABA/-AAA) and lower layer (UDP for both

protocols). The results are shown in Table 5.

Table 5 show that for the cases mobile client

moving at high speed, the efficiency of QUIC-ABA

is about 76% higher than that of QUIC-AAA and 17%

higher than that of UDP-ABA; at slow speeds, the

efficiency of QUIC-ABA is 17% higher than that of

QUIC-AAA, and about 48% higher than UDP-ABA.

Note that the data shown for UDP-ABA scheme in

Table 5 is an upper bound pps, since we did not take

those loss packets into account. So the actual packet

received rate at receiver side for UDP-ABA should be

lower than it. The results show that the proposed

scheme QUIC-ABA out performs the others.

Table 4: packet receive rate (sender rate= unbounded).

scheme speed pps@receiver

QUIC-ABA fast 181.82

QUIC-AAA 101.91

QUIC-ABA slow 176.45

QUIC-AAA 134.35

Table 5: comparisons of packet receive rate at different

layers (sender rate= unbounded).

scheme

speed pps@receiver

UDP

QUIC-ABA fast

181.82 389.48

QUIC-AAA

101.91 220.77

UDP-ABA

155.43 244.53

QUIC-ABA slow

176.45 385.89

QUIC-AAA

134.36 293.65

UDP-ABA

119.43 313.54

4 CONCLUSIONS

In this study we propose a novel dual-connection Wi-

Fi mesh system framework for efficient mobile data

transmission, in which we combine multiple Wi-Fi

transmission mechanism to QUIC protocol to create

a new type of multipath network transmission system.

This system aims to reduce the adverse effects of

physical environmental changes and wireless access

point (AP) device transitions, achieving a low-cost,

high-speed, highly mobile, and secure network

environment in specific settings. The proposed

scheme leverages the encryption and security features

of the QUIC protocol to protect data privacy while

supporting the needs of high-mobility applications.

We conduct experiments to evaluate the proposed

scheme under different configurations including

number of Wi-Fi interfaces, Wi-Fi SSID assignment,

sender data generation rates, and client moving

speeds. Experimental results show that the proposed

scheme QUIC-ABA out performs the others.

REFERENCES

Pahlavan, K., & Krishnamurthy, P. (2021). Evolution and

Impact of Wi-Fi Technology and Applications: A

Historical Perspective. International Journal of

Wireless Information Networks, 28(3), 3-19.

https://doi.org/10.1007/s10776-020-00501-8

IoTBDS 2025 - 10th International Conference on Internet of Things, Big Data and Security

454

Qureshi, I. A., & Asghar, S. (2023). A Systematic Review

of the IEEE-802.11 Standard’s Enhancements and

Limitations. Wireless Personal Communications, 131,

2539-2572. https://doi.org/10.1007/s11277-023-

10553-7

Liao, R., Bellalta, B., Barcelo, J., Valls, V., & Miquel, O.

(2013). Performance analysis of IEEE 802.11ac

wireless backhaul networks in saturated conditions.

EURASIP Journal on Wireless Communications and

Networking, 226 (2013). https://doi.org/10.1186/1687-

1499-2013-226

Deng, D.-J., Lin, Y.-P., Yang, X., Zhu, J., Li, Y.-B., Luo,

J., & Chen, K.-C. (2017). IEEE 802.11ax: Highly

Efficient WLANs for Intelligent Information

Infrastructure. IEEE Communications Magazine,

55(12), 172-178. https://doi.org/10.1109/

MCOM.2017.1700285

Polese, M., Chiariotti, F., Bonetto, E., Rigotto, F., Zanella,

A., & Zorzi, M. (2019). A Survey on Recent Advances

in Transport Layer Protocols. IEEE Communications

Surveys & Tutorials.

https://doi.org/10.1109/COMST.2019.2932905

Kulkarni, S., & Agrawal, P. (2014). Analysis of TCP

Performance in Data Center Networks. Springer Briefs

in Electrical and Computer Engineering. Springer, New

York. ISBN 978-1-4614-7861-4. DOI: 10.1007/978-1-

4614-7861-4.

Vedavathi, T., Karthick, R., Selvan, R. S., & Meenalochini,

P. (2020). Data Communication and Networking

Concepts in User Datagram Protocol (UDP).

International Journal of Recent Technology and

Engineering (IJRTE), 8(5), 2765-2768. DOI:

10.35940/ijrte.D8758.018520.

Langley, A., Riddoch, A., Wilk, A., Vicente, A., Krasic, C.,

Zhang, D., Yang, F., Kouranov, F., Swett, I., Iyengar,

J., Bailey, J., Dorfman, J., Roskind, J., Kulik, J., Westin,

P., Tenneti, R., Shade, R., Hamilton, R., Vasiliev, V.,

Chang, W., & Shi, Z. (2017). The QUIC Transport

Protocol: Design and Internet-Scale Deployment. In

Proceedings of SIGCOMM '17, Los Angeles, CA,

USA, August 21–25, 2017.

Qin, X., Chen, X., Huang, W., Xie, Y., & Zhang, Y. (2022).

Measurement and Analysis: Does QUIC Outperform

TCP? In Proceedings of 18th International Conference

on Mobility, Sensing and Networking (MSN). DOI:

10.1109/MSN57253.2022.00154

A Novel Wi-Fi Mesh Network Framework for Efﬁcient Mobile Data Transmission

455