Dezheng Kong

Mr. Dezheng Kong

Department of Information and Communication Systems, Okayama University, Okayama, Japan

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 March 2024 in Sensors
Verified
Journal Article
Open access
PDF file
BIN WUBIN WU
BIN WU
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
SRSujan Chandra Roy
Sujan Chandra Roy
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
DKDezheng Kong
Dezheng Kong
FSFang Shihao
Fang Shihao
With the wide applications of the Internet of Things (IoT) in smart home systems, IEEE 802.11n Wireless Local Area Networks (WLANs) have become a frequently chosen communication technology due to their adaptability and affordability. In a high-density network of devices such as the smart home scenerio, a host often meets interferences from other devices and unequal Received Signal Strength (RSS) from Access Points (APs). This results in throughput unfairness/insufficiency problems between hosts communicating concurrently in WLAN. Previously, we have studied the throughput request satisfaction method to address this problem. It calculates the target throughput from measured single and concurrent throughputs of hosts and controls the actual throughput at this target one by applying traffic shaping at the AP. However, the insufficiency problem of maximizing the throughput is not solved due to interferences from other hosts. In this paper, we present an extension of the throughput request satisfaction method to maximize the throughput of a high-priority host under concurrent communications. It recalculates the target throughput to increase the actual throughput as much as possible while the other hosts satisfy the least throughput. For evaluations, we conduct experiments using the test-bed system with Raspberry Pi as the AP devices in several topologies in indoor environments. The results confirm the effectiveness of our proposal.

7 February 2026 in Sensors
Verified
Journal Article
Open access
PDF file
DKDezheng Kong
Dezheng Kong
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
HKHtoo Htoo Sandi Kyaw
Htoo Htoo Sandi Kyaw
IKI Nyoman Darma Kotama
I Nyoman Darma Kotama
ZZZhu Zihao
Zhu Zihao
ARAlfiandi Aulia Rahmadani
Alfiandi Aulia Rahmadani
Nowadays, Internet of Things (IoT) application systems play an essential role in smart cities, industry, healthcare, agriculture, and smart homes. For non-expert users, designing and implementing IoT application systems remains challenging, especially when configuring sensors, edge devices, and server platforms. To support configuration tasks of IoT application systems, we have developed an AI-based setup assistance tool. However, AI models still fail to reliably support newly released or previously unseen devices, sometimes producing incomplete or erroneous outputs that may lead to configuration failures. Incorporating their technical-document information into Retrieval-Augmented Generation (RAG) is an effective way to supplement AI knowledge and improve reliability. In this paper, we propose a generative AI-based technical data extraction tool to address the challenges. It extracts essential technical information using the schema-based extraction from given PDF or HTML datasheets and converts it into a structured format suitable for AI-supported configurations. A local vector database is used to enable semantic similarity retrieval and provide document-grounded evidence for RAG-based answering, ensuring consistent support for previously unseen IoT devices. For evaluations, we applied the proposal to several sensor and device datasheets and compared extracted specifications with ground-truth values to measure accuracy and completeness. Then, we compared end-to-end configuration QA reliability against a commercial baseline (ChatPDF) using the golden benchmark. The results show that the proposed tool reliably acquires key specifications and significantly improves end-to-end configuration QA reliability. Across 960 golden QA pairs, the proposed method improves Recall from 0.636 to 0.926 and Accuracy from 0.595 to 0.807 compared with ChatPDF.

13 July 2023 in Sensors
Verified
Journal Article
Open access
PDF file
FAFatema Akhter
Fatema Akhter
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
EHEi Ei Htet
Ei Ei Htet
BIN WUBIN WU
BIN WU
DKDezheng Kong
Dezheng Kong
SFShihao Fang
Shihao Fang
The IEEE 802.11 wireless local-area network (WLAN) has been deployed around the globe as a major Internet access medium due to its low cost and high flexibility and capacity. Unfortunately, dense wireless networks can suffer from poor performance due to high levels of radio interference resulting from adjoining access points (APs). To address this problem, we studied the AP transmission power optimization method, which selects the maximum or minimum power supplied to each AP so that the average signal-to-interference ratio (SIR) among the concurrently communicating APs is maximized.However, this method requires measurements of receiving signal strength (RSS) under all the possible combinations of powers. It may need intolerable loads and time as the number of APs increases. It also only considers the use of channel bonding (CB), although non-CB sometimes achieves higher performance under high levels of interference. In this paper, we present an AP interface setup optimization method using the throughput estimation model for concurrently communicating APs. The proposed method selects CB or non-CB in addition to the maximum or minimum power for each AP. This model approach avoids expensive costs of RSS measurements under a number of combinations. To estimate the RSS at an AP from another AP or a host, the model needs the distance and the obstacles between them, such as walls. Then, by calculating the estimated RSS with the model and calculating the SIR from them, the AP interface setups for a lot of APs in a large-scale wireless network can be optimized on a computer in a very short time. For evaluation, we conducted extensive experiments using Raspberry Pi for APs and Linux PCs for hosts under 12 network topologies in three buildings at Okayama University, Japan, and Jatiya Kabi Kazi Nazrul Islam University, Bangladesh. The results confirm that the proposed method selects the best AP interface setup with the highest total throughput in any topology.

7 August 2024 in Symmetry
Verified
Journal Article
Open access
PDF file
BIN WUBIN WU
BIN WU
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
DKDezheng Kong
Dezheng Kong
XWXuan Wang
Xuan Wang
TSTaishiro Seto
Taishiro Seto
Yu-Cheng FanYu-Cheng Fan
Yu-Cheng Fan
Wireless Local-Area Networks (WLANs), as a popular internet access solution, are widely used in numerous places, including enterprises, campuses, and public venues. As the number of devices increases, large-scale deployments will cause the problem of dense wireless networks, including a lot of energy consumption. Thus, the optimization of energy-efficient wireless AP devices has become a focal point of attention. To reduce energy consumption, we have proposed the active access-point (AP) configuration algorithm for WLANs using APs with a dual interface. This uses the greedy algorithm combined with the local search optimization method to find the minimum number of activated APs while satisfying the minimum throughput constraint. However, the previous algorithm basically satisfies only the average throughput among the multiple hosts associated with one AP, wherein some hosts may not reach the required one. In this paper, to overcome this limitation, we propose an enhanced active AP configuration algorithm by incorporating the throughput request satisfaction method that controls the actual throughput at the target value (target throughput) for every host by applying traffic shaping. The target throughput is calculated from the single and concurrent communicating throughput of each host based on channel occupancy time. The minimum throughput constraint will be iteratively adjusted to obtain the required target throughput and achieve the fair throughput allocation. For evaluations, we conducted simulations using the WIMNET simulator and experiments using the testbed system with a Raspberry Pi 4B for APs in four topology cases with five APs and ten hosts. The results show that the proposed method always achieved the required minimum throughput in simulations as well as in experiments, while minimizing the number of active APs. Thus, the validity and effectiveness of our proposal were confirmed.

19 August 2026 in Future Internet
Verified
Journal Article
Open access
PDF file
DKDezheng Kong
Dezheng Kong
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
HKHtoo Htoo Sandi Kyaw
Htoo Htoo Sandi Kyaw
IKI Nyoman Darma Kotama
I Nyoman Darma Kotama
ZZZhu Zihao
Zhu Zihao
Internet of Things (IoT) application systems are increasingly adopted in factories, shops, offices, and governments. However, building such systems using various devices and modules remains difficult for non-experts, because they must confirm specifications, communication interfaces, voltage ranges, and operating conditions from technical datasheets before connecting devices. In previous studies, we have explored a generative AI-based answering tool for datasheet inquiry using Retrieval-Augmented Generation (RAG) for technical guidance of IoT application system deployment. However, the adopted top-kRAG pipeline often retrieves multiple related text chunks, which can cause the AI to confuse technically different specifications, such as power output voltage, signal output voltage, and input voltage range, and produce inaccurate answers. In addition, the AI may generate a hallucinated answer if the datasheet does not provide sufficient source information. In this paper, we propose an AI answer-validation method using agentic RAG for datasheet inquiry for IoT application system deployment. The method organizes datasheet information into structured specification data, including device models, field types, values, units, conditions, and source information. For question-answering, the agent coordinates structured fact query, top-k text retrieval, source checking, and rule-based compatibility comparison according to the question type. Instead of fully relying on the LLM to interpret retrieved chunks, this method adopts structured specifications and deterministic source checks before accepting the final answer. For evaluation, we constructed a dataset from 20 IoT datasheets, including 1000 question-answering tasks with three difficulty levels. Compared with conventional top-k RAG, the proposed method improved the correct answer rate from 0.686 to 0.958 for easy questions, from 0.549 to 0.969 for medium questions, and from 0.273 to 0.613 for hard questions, which confirms the effectiveness of the proposed method.

2 April 2025 in Technologies
Verified
Journal Article
Open access
PDF file
DKDezheng Kong
Dezheng Kong
Nobuo FunabikiNobuo Funabiki
Nobuo Funabiki
FSFang Shihao
Fang Shihao
NNNoprianto Noprianto
Noprianto Noprianto
MOMitsuhiro Okayasu
Mitsuhiro Okayasu
PPPradini Puspitaningayu
Pradini Puspitaningayu
For decades, analog dial gauges have been essential for measuring and monitoring data at various industrial instruments including production machines and laboratory equipment. Among them, we focus on the instrument for creep test in a mechanical engineering laboratory, which evaluates material strength under sustained stress. Manual reading of gauges imposes significant labor demands, especially in long-duration tests. This burden further increases under low-lighting environments, where poor visibility can lead to misreading data points, potentially compromising the accuracy of test results. In this paper, to address the challenges, we implement a creep test assisting system that possesses the following features: (1) to save the installation cost, a web camera and Raspberry Pi are employed to capture images of the dial gauge and automate the needle reading by image processing in real time, (2) to ensure reliability under low-lighting environments, a smart lighting mechanism is integrated to turn on a supplementary light when the dial gauge is not clearly visible, and (3) to allow a user to stay in a distant place from the instrument during a creep test, material break is detected and the corresponding message is notified to a laboratory staff using LINE automatically. For evaluations, we install the implemented system into a material strength measuring instrument at Okayama University, Japan, and confirm the effectiveness and accuracy through conducting experiments under various lighting conditions.