International Journal of Wireless and Mobile Computing (57 papers in press)

Regular Issues

• Multi-objective workflow scheduling in the cloud environment based on NSGA-II  
  by Tingting Dong, Chuangbai Xiao 
  Abstract: The emergence of cloud computing offers a novel perspective to solve large-scale computing problems. Workflow scheduling is a major problem in the cloud environment, and parallelism and dependency are two important characteristics of tasks in a workflow, which increases the complexity of problem. Workflow scheduling is also a multi-objective scheduling problem, and task execution time and cost are the two extremely significant goals for users and providers in the cloud environment. Existing heuristic algorithms are popular, but they lack of robustness and need to be revised when the problem statement changes. Evolutionary algorithms have a complete algorithm system, which is widely used in the multi-objective scheduling problem. In this paper, Nondominated Sorting Genetic Algorithm-II (NSGA-II) is utilized to solve the workflow scheduling problem aiming at minimizing the task execution time and cost. Some real-world workflows are used to make simulation expriments, and comparative simulations with genetic algorithm are given. Results show that NSGA-II is effective for the workflow scheduling.
  Keywords: cloud computing; workflow scheduling; non-dominated sorting genetic algorithm-II; multi-objective scheduling.
  
  
• Enhancing artificial bee colony algorithm with depth-first search and direction information  
  by Xinyu Zhou, Hao Tang, Shuixiu Wu, Mingwen Wang 
  Abstract: In recent years, artificial bee colony (ABC) algorithm has been criticized for its solution search equation, which makes the search capability bias towards exploration at the expense of exploitation. To solve the defect, many improved ABC variants have been proposed aiming to use the elite individuals. Although these related works have shown effectiveness, they rarely take the factor of search direction into account. In fact, the search direction has an important role in determining the performance of ABC. Thus, in this work, we are motivated to investigate how to combine the idea of using the elite individuals with the search direction, and a new ABC variant, called DDABC, is designed. In the DDABC, the depth-first search (DFS) mechanism and direction information learning (DIL) mechanism are introduced, and the former mechanism is to allocate more computation resources to the elite individuals, while the latter mechanism aims to adapt the search to the promising directions. To verify the effectiveness of the DDABC, experiments are carried out on 22 classic test functions, and three relative ABC variants are included as the competitors. The comparison results show the competitive performance of our approach.
  Keywords: artificial bee colony; exploration and exploitation; depth-first search; direction information learning.
  
  
• Machine learning-based approach for the detection of phishing websites  
  by Yaqin Wang, Jingsha He, Nafei Zhu 
  Abstract: Compared with traditional forms of crime, cyber-attacks and cyber-crimes have removed the limitation on distance and speed. With very low cost, phishing is a very effective way of launching network attacks with the purpose of obtaining sensitive information about users, such as username, password and payment voucher, through counterfeiting regular websites so as to steal users private information and personal property using the obtained information. Both the trust that internet users have and the development of the internet itself can be affected by this kind of attack, making it imperative to detect this type of attack. Many methods have been proposed for the detection of phishing websites in the literature in recent years based on techniques ranging from conventional classifiers to complex hybrid classifiers. Meanwhile, although convolutional neural networks (CNNs) can achieve very high accuracy in classification tasks, not much research has been done on the use of CNNs for the detection of phishing websites. This paper proposes a CNN-based scheme for the detection of phishing websites in which four dimensions of the features of phishing websites are defined and CNN is used to extract local features. The proposed CNN-based scheme is compared with several machine learning-based methods on the effectiveness of detecting phishing websites, which shows that the proposed scheme can achieve the accuracy rate of 97.39% and is better than the other classification methods in terms of accuracy, recall and F1-score.
  Keywords: convolutional neural network; classification; machine learning; phishing website detection.
  
  
• Cold-start recommendation algorithm based on user preference estimation  
  by Biao Cai, Jiahui Xin, Xu Ou 
  Abstract: In order to improve the dilemma of collaborative filtering in the face of cold start and achieve a better balance between accuracy and diversity, this paper considers the influence of user characteristics on recommendation results and proposes a Preference Estimation Network (PEN) based on maximum likelihood. PEN uses the user's characteristic information to estimate the user's preference information, and represents the user's preference vector with the item's label system. On this basis, PEN-Rec, an improved version of the traditional recommendation algorithm based on preference vector estimation and particle swarm optimisation, is proposed. Finally, the PEN-Rec algorithm is compared with the benchmark algorithm on six public evaluation indicators using open datasets, and the experimental results show that the accuracy, diversity and novelty of the PEN-Rec algorithm are all improved.
  Keywords: recommendation; feature impact; preference estimation; label vector.
  
  
• Link prediction with Fusion of DeepWalk and node structural information  
  by Xinhui Xiang, Biao Cai, Yunfen Luo 
  Abstract: The existing link prediction algorithms are mainly based on structural information or network embedding, but minimal research has been conducted on the fusion of these algorithms. It is found that the structure-based algorithms have high accuracy, but the complexity is higher owing to the introduction of high-order structural information while the network embedding algorithms have low complexity, but because the structural information of the node is not fully used, the accuracy is not as good as some structure-based algorithms. Therefore, by combining the structural attributes of nodes and the degree of convergence between node pairs, this paper proposes two new improved similarity algorithms the similarity algorithm based on edge-degree DeepWalk cosine (EDDWC) and the similarity algorithm based on preferential attachment mechanism DeepWalk cosine (PADWC). Experiments show that the performances of the proposed algorithms are greatly improved over that of the DeepWalk algorithm, and they are also better than other link prediction algorithms.
  Keywords: link prediction; DeepWalk; edge-degree; preferential attachment mechanism; cosine similarity.
  
  
• Hop count, ETX and energy selection based objective function for image data transmission over 6LoWPAN in IoT  
  by Archana Bhat, Geetha V 
  Abstract: Internet of things (IoT) is technology that connects millions of things to the internet for collecting data and controlling things. 6LoWPAN looks promising for future IoT networks as it works with IPv6, which is essential to address millions of things. However, as the 6LoWPAN devices are resource constrained with payload constraint at the data link layer, it needs efficient mechanisms to send packets over IEEE 802.15.4 MAC layer. The challenge increases when the sensors used in the devices are camera or audio recordings. Multimedia data transmission over 6LoWPAN is great challenge, and this paper addresses the same with respect to selection of Objective Function (OF) for multimedia data traffic. A new hop count, ETX and energy selection based OF is proposed in this work. The proposed technique is compared with existing OF, and the simulation results shows that the proposed technique provides better performance.
  Keywords: 6LoWPAN; objective function; IPv6; multimedia; RPL; IEEE 802.15.4.
  
  
• Poor and rich squirrel algorithm-based Deep Maxout network for credit card fraud detection  
  by Annu Paul, Varghese Paul 
  Abstract: This paper proposes a Poor and Rich Squirrel Algorithm (PRSA)-based Deep Maxout network to find fraud data transactions in the credit card system. Initially, input transaction data is passed to the data transformation phase, transforming data using Yeo-Johnson (YJ) transformation. Then, the feature selection procedure is done by the Fisher score for creating the unique and significant features. Next, based on the selected textures, the data augmentation mechanism is done using the oversampling model. At last, the fraud detection is carried out by the Deep Maxout network, which is trained by the proposed PRSA optimisation algorithm, derived by integrating Poor and Rich Optimisation (PRO) and Squirrel Search Algorithm (SSA). The integration of parametric features of the PRSA algorithm trained the classifier to update weights to generate the best solution by considering fitness measures. The proposed method achieved the best accuracy, sensitivity, and specificity measures of 0.96, 0.95, and 0.94, respectively.
  Keywords: credit card; deep learning; fraud detection; data augmentation; data transformation.
  
  
• A study on dual-sense broadband circularly polarised monopole antenna for UWB applications  
  by Umesh Singh, Kalyan Mondal, Rajesh Mishra 
  Abstract: The proposed work is designed with the embedded of stubs and Parasitic Strips (PSs) under the radiator. An FR4 substrate is used to design the antenna (?_r= 4.4, h = 1.6 mm). The overall size of the antenna is 0.8?_0
  Keywords: monopole; CP; dual-sense; stubs; ARBW; satellite;.
  
  
• A novel trust-based approach for intrusion detection architecture in wireless sensor networks  
  by Mr. Jeelani, Kishan Pal Singh, Aasim Zafar 
  Abstract: Wireless sensor networks (WSNs) is a new technology that can be used to monitor the environment. Because sensor nodes in wireless sensor networks are installed in an open environment, they are more vulnerable to attacks. The sensor network lifetime improvement is dependent on minimum energy use. Protection is also a major concern when it comes to designing protocols for multi-hop secure routing. The results based on trust have proven to be more effective in addressing malicious node attacks. In this article, we propose a novel trust-based approach for intrusion detection architecture (IDA) in a wireless sensor network that is called the trust-based approach for varying nodes with energy (TBNE) model. TBNE finds the misbehaving nodes in the network. The structure is based on the trust model for secure communication in WSN and improves the performance of nodes. The simulation has been done with QualNet 5.0 simulator.
  Keywords: wireless sensor network; throughput; packet delivery ratio.
  
  
• Data sharing with privacy protection based on blockchain and federated learning in edge computing enabled IoT  
  by Shiqiang Zhang, Zhenhu Ning 
  Abstract: Data sharing of Internet of things devices is a powerful means and technology to break the data island in the era of big data. However, frequent privacy leaks indicate that privacy protection has become one of the most urgent problems in data sharing. The existing data sharing schemes usually provide data to the data demanders through access control authorisation through a third-party organization. This way can protect the privacy of data to a certain extent. But the biggest problem is that the data owner will lose control of the data, which increases the risk of privacy disclosure. In this paper, we proposed a new data sharing scheme based on blockchain and federated learning. The data sharing problem is transformed into a machine learning problem. The IoT devices train the model locally and use differential privacy technology to avoid privacy leakage, and ensures its security through the blockchain network aggregation model.
  Keywords: data sharing; blockchain; federated learning; differential privacy; edge computing; IoT.
  
  
• A survey of lung nodule computer-aided diagnostic system based on deep learning  
  by Tongyuan Huang, Yuling Yang 
  Abstract: With the development of machine learning, especially deep learning, the research of pulmonary nodules based on deep learning has made great progress, which has important theoretical research significance and practical application value. Therefore, it is necessary to summarise the latest research in order to provide some reference for researchers in this field. In this paper, the related research, typical methods and processes in the field of pulmonary nodules are analysed and summarised in detail. Firstly, the background knowledge in the field of pulmonary nodules is introduced. Secondly, the commonly used data sets and evaluation indexes are summarised and analysed. Then, the computer-aided diagnostic system related processes and key sub problems are summarised and analysed. Finally, the development trend and conclusion of pulmonary nodule computer-aided diagnostic system are prospected.
  Keywords: machine learning; deep learning; pulmonary nodule; CAD system.
  
  
• Technology adoption of enablers of 5G networks for m-learning: an analysis with interpretive structural modelling and MICMAC  
  by L. Kala, Hameed T. A. Shahul, V.R. Pramod 
  Abstract: Mobile learning (m-learning) is one of the real-time applications of 5G technology with an impulsive future. COVID-19 pandemic enhanced the adoption of m-learning over wireless networks by facilitating continued formal education or work from home. This research aims to analyse enablers of 5G networks that enhance real-time m-learning by applying Interpretive Structural Modelling (ISM), a set-theory-based structural modelling method widely employed in many engineering and technology related research fields. Data was collected through questionnaire-based information gathering and from one-to-one discussions with experts. Modelling was performed to identify the correlations among system parameters through a hierarchically structured model. Further, the enablers were classified into different clusters based on their driving powers and dependency with MICMAC analysis, by which the results were validated. The study shows that enablers of 5G will undoubtedly support and uphold the system performance for future real-time scenarios of m-learning by eliminating all the inhibiting parameters of former 4G wireless networks.
  Keywords: 5G; wireless networks; enablers; mobile learning; ISM; MICMAC; driving power dependence.
  
  
• Two-phase approach for the detection and isolation of black hole attack in mobile ad hoc network  
  by Pankaj Khuresha, Sonal Sood, Mandeep Sandhu, Anurag Dixit 
  Abstract: A mobile ad hoc network (MANET) is an infrastructure-less network in which no central controller is present and nodes can communicate with each other independently. Owing to unique nature of the network, malicious nodes can enter the network which triggers various types of attack. The black hole is the attack in which the malicious node does not forward any packets and all the packets will be dropped in the network. In this research work, an approach is proposed for the detection and isolation of black hole attack in MANET. The proposed approach works in two phases: in the first phase the malicious node will be detected and in the second phase the malicious node will be isolated from the network. The proposed methodology is implemented in network simulator version 2 and results are analysed in terms of throughput, delay and packet loss.
  Keywords: MANET; black hole; malicious nodes; clustering; trust.
  
  
• Optimised recurrent neural network based localisation in wireless sensor networks: a composite approach  
  by Shivakumar Kagi, Basavaraj S. Mathapati 
  Abstract: Localisation is one of the key techniques in the wireless sensor network. The location estimation methods can be classified into target/source localisation and node self-localisation. There are several challenges in some special scenarios. Therefore, the anchor node-based distance estimation scheme is used in this research work. In the anchor-based localisation technique, the unknown node uses the position of the anchor node to estimate its location. The trained Recurrent Neural Network (RNN) with the extracted Angle Of Arrival (AoA) and RSSI features of the anchor node and the estimated nodes makes the localisation of the unknown node more precise. Further, to lessen the localisation errors in RNN, its weights are fine-tuned by an Improved Whale optimisation Algorithm (IWOA).
  Keywords: WSN; node localisation; AoA and RSSI based feature computation; RNN; IWOA.
  
  
• LTE 2100 MHz band half-wave two element rectifier array for wireless electromagnetic energy harvesting  
  by Pradeep Chandrakant Dhanawade, Shivajirao M. Sangale 
  Abstract: In this manuscript, a two-element half-wave rectenna array for wireless energy harvesting from LTE2100 MHz band is presented. The 2100 MHz band is chosen based on the spectrum survey in the locality. An outdoor peak power of -15.8 dBm is sensed using a 1.5 dBi gain wideband antenna and spectrum analyser. A half-wave rectifier circuit using two different Schottky diodes and a capacitor filter is developed and connected in mirror image form. The proposed structure combines the direct current power of individual elements using two series capacitors improving the rectenna efficiency. The reported full-wave rectifier array has 19.95% and 63.01% radio-frequency to direct current efficiency for conventional and high-performance Schottky diodes respectively. A detailed analysis of major design parameters have been performed and presented in the manuscript which will help researchers to choose a suitable operating band and design components for rectenna design. The presented half-wave-rectifier rectenna has a comparable conversion efficiency with the full-wave-rectifier rectennas resulting in improved throughput wireless energy harvesting systems.
  Keywords: rectifier; rectenna; Schottky diode; wireless energy harvesting; rectenna array.
  
  
• Signal strength and energy based efficient AODV routing algorithm in MANET  
  by Priyanks Pandey, Raghuraj Singh 
  Abstract: In recent years, Mobile Ad Hoc Network (MANET) has become one of the most popular research areas in the wireless networking domain. However, one of the major challenges remains to develop an efficient routing algorithm which depicts par excellence performance on all performance parameters even under highly dynamic network. Ad Hoc On-Demand Distance Vector (AODV) is a generalized routing protocol which establishes routes to destinations on demand in MANET environment and supports unicast as well as multicast routing. Many enhancements have also been proposed in AODV from time to time. These enhancements are based on various features which define a specific environment. But, these enhancements do not perform well on all considered performance metrics such as packet delivery ratio, delay, normalized routing load and throughput in highly dynamic network environment. In this paper, we have proposed an Enhanced version of AODV, namely (ENAODV) algorithm considering two important and additional stability parameters i.e. energy and signal strength along with hop count and sequence number in route selection process. Algorithm has been simulated using NS2 simulator and evaluated under different network conditions with varying maximum speed. Performance of the algorithm has been evaluated to be better on all parameters like throughput, normalized routing load, packet delivery ratio, control overhead and end to end delay than the AODV algorithm.
  Keywords: MANET; signal strength; RWP; AODV.
  
  
• QoS-based handover approach for 5G mobile communication system  
  by Amina Gharsallah, Nouri Omheni, Faouzi Zarai, Mahmoud Neji 
  Abstract: 5G mobile communication systems are in-depth fusions of multi-radio access technologies characterised by frequent handover between cells. Handover management is a particularly challenging issue for 5G networks development. In this article, a novel optimised handover framework is proposed to find the optimal network to connect with a good quality of service in accordance with the users preferences. This framework is based on an extension of IEEE 802.21 standard with new components and new service primitives for seamless handover. Moreover, the proposed vertical handover process is based on an adaptive heuristic model aimed at achieving an optimised network during the decision-making stage. Simulation results demonstrate that, compared to other existing works, the proposed framework is capable of selecting the best network candidate accurately based on the quality of service requirements of the application, network conditions, mobile terminal conditions and user preferences. It significantly reduces the handover delay, handover blocking probability and packet loss rate.
  Keywords: 5G mobile network; ultra-dense network; media independent handover; vertical handover optimisation; fast handover.
  
  
• A feature fusion pedestrian detection algorithm  
  by Nan Xiang, Lu Wang, Xiaoxia Ma, Chongliu Jia, Yuemou Jian, Lifang Zhu 
  Abstract: When pedestrians are in different angles and positions, The feature extraction and fusion capabilities are often limited of YOLO series model. Aimed at this problem, we propose an improved feature fusion pedestrian detection algorithm YOLO-SCr. To enhance the ability of cross-scale feature extraction and detection speed, we reconstruct the network structure of the YOLO algorithm in the backbone part and convolution layer part, respectively. Then, to strengthen the feature fusion ability of pedestrians at different scales ,we introduce the spatial pyramid pooling module and shuffle & CBAM(Convolutional Block Attention Module) attention mechanisms in different positions before YOLO layers. The experimental results show that compared with the detection algorithm such as YOLOv3, YOLO-SCr can performance effectively improve the detection accuracy , Recall and speed.
  Keywords: YOLO series ; feature extraction ; feature fusion ;spatial pyramid pooling; pedestrian detection ; shuffle & CBAM attention;.
  
  
• Research on a laser cutting path planning method based on improved ant colony optimisation  
  by Naigong Yu, Qiao Xu, Zhen Zhang 
  Abstract: Laser cutting path planning for fabric patterns is critical to cutting efficiency. The ant colony optimisation algorithm commonly used in this field is constrained by the complete cutting and cannot plan a true global optimal path, resulting in large empty strokes. To solve this problem, this paper proposes an ant colony optimisation method based on virtual segmentation of multiple feature points for path planning of laser cutting. The method first changes the feature point selection strategy of traditional ant colony optimisation and increases the number of feature points in a single pattern. Then the single closed pattern is virtually divided into multiple open contours. Finally, the optimal cutting path is planned based on the solution of the travelling salesman problem. Experiments show that the cutting planning path obtained by the proposed method has a higher degree of compression on the idle stroke and significantly improves the laser cutting efficiency.
  Keywords: laser cutting; path planning; ant colony optimisation; virtual segmentation.
  
  
• Two novel blind CFO estimation techniques for CP-OFDM  
  by Mohammadreza Janbazi Roudsari, Javad Kazemitabar, Hossein Miar-Naeimi 
  Abstract: In this paper, two new cyclic prefix (CP) based blind carrier frequency offset (CFO) estimation methods for orthogonal frequency division multiplexing (OFDM) transmission over multipath channels are proposed. In doing so, we first estimate the maximum delay of the fading channel. We borrow the concept of remodulation introduced in earlier works and use the repetitive structure of CP to calculate a maximum-likelihood based measure. In the first proposed method we use particle swarm optimisation aided search on all possible samples to find the optimal set. This technique provides performance improvement at the expense of more complexity. Then, in a second proposed method, we average over the optimal set of samples to estimate CFO. The second technique provides a major improvement over previous works while offering less complexity. Simulation results corroborate that both our proposed methods significantly decrease the mean square error.
  Keywords: orthogonal frequency division multiplexing; carrier frequency offset; cyclic prefix.
  
  
• Towards energy-efficient 5G heterogeneous networks through dynamic small cell zoom and sleep control algorithm  
  by Janani Natarajan, Rebekka B 
  Abstract: The tremendously escalating mobile traffic and bandwidth hungry applications is challenging the network operators to provide guaranteed quality of service (QoS) over wider coverage and effective network resource usage. One of the effective solutions is heterogeneous network (HetNet) comprising an overlay of small cells (SCs) within a macrocell coverage. For enhancement in network energy efficiency (EE), we propose a joint small cell zoom and sleep strategy. The small cell zoom technique involves load-aware adaptive power control of the SCs for optimum network power consumption through lower SBS use together with appropriate user load balance. The small cell sleep method switches the SBSs with higher interference to sleep mode, thereby improving the network capacity as well as power saving. Simulation results show an EE improvement of the proposed sleep and zoom scheme by 25%, 26% and 28%, respectively, compared with three similar benchmark schemes in the literature.
  Keywords: heterogeneous networks; small cells; energy efficiency; small cell zoom; small cell sleep; adaptive power control.
  
  
• Deep reinforcement learning multi-robot cooperative scheduling based on service entity network  
  by Xueguang Jin, Chengrui Wu, Yan Yan, Yingli Liu 
  Abstract: Multi-robots are increasingly deployed with the development of automation in agriculture, industry, and warehousing logistics. With the help of CPS virtualisation technology, services or tasks can be decomposed into a network with capability or function entity nodes and edges connecting nodes. In this paper, the service entity network is extended with human, robot, and IT resources as a task-decomposed network with public entities, private entities, and links. Based on the service entity network virtualisation architecture, it is possible to form a global service entity network corresponding to the correlated tasks. Meanwhile, deep reinforcement learning multi-robot cooperative scheduling based on a service entity network framework is studied, which makes it possible to jointly optimise the deployment of multi-robot tasks with multi-service entity networks. The results show that the model based on the artificial intelligence virtualisation architecture achieves a better performance.
  Keywords: service entity network; virtualisation technology; multi-robot cooperative scheduling.
  
  
• SBER: Stable and Balance Energy Routing Protocol to Enhance the Stability and Energy for WBANs  
  by Sara Raed, Salah Abdulghani Alabady 
  Abstract: Stability and reduced energy consumption are essential in the design requirements of Wireless Body Area Network (WBAN) routing protocols. For instance, many energy-efficient routing protocol solutions have been suggested for WBANs; however, the significant feature of stability in these existing solutions has not been effectively addressed. In this paper, we propose a Stable and Balance Energy Routing (SBER) protocol to improve the stability period and manage the limited power of the WBAN network efficiently. SBER consists of two solutions, namely, the next-hop node selection and adding awareness to the transmission of control packets techniques. For analysis of the performance of the SBER protocol, MATLAB has been used. The average improvements rate of the SBER in terms of network residual energy over ERRS, M-ATTEMPT, and SIMPL protocols are 35%, 52%, and 100% respectively, which proves SBER to be a more efficient and reliable approach for WBANs.
  Keywords: WBANs; stability period; routing protocol; SBER; ERRS; M-ATTEMPT; SIMPL.
  
  
• Research on fire alarm system of the intelligent building based on information fusion  
  by Sun Xuejing 
  Abstract: In order to effectively reduce the hazards caused by fire and improve the accuracy of fire alarm systems, this paper proposes to use STM32 microcontroller as the control core, use the communication method of Zigbee wireless communication technology combined with CAN bus technology, apply the QPSO-BP neural network algorithm based on multi-sensor information fusion method to fire alarm judgment, and use the fire protection partition in the building as the basis for the distributed intelligent building fire alarm system. The results show that the distributed intelligent building fire alarm system designed in this paper meets the design requirements of the system while fully considering the economic benefits and makes up for the shortcomings of the traditional fire alarm system. The algorithm output results are accurate and reliable, providing a reference for the design of building fire alarm systems.
  Keywords: intelligent building fire alarm; information fusion; QPSO-BP neural network algorithm; Zigbee technology.
  
  
• A hybrid meta-heuristic algorithm to detect malicious activity based on dynamic ON VANET environmental information  
  by Gagan Preet Kour Marwah, Anuj Jain 
  Abstract: VANET has the characteristics of self-organisation, rapid topology changes, and frequent link disconnection that perhaps led to challenging issues. In order to mitigate these issues, a highly effective technology is required; therefore, this work has adopted a Hybrid Firefly Optimisation Algorithm (FOA) and a Whale Optimisation Algorithm (WOA) named as HFWOA-VANET. The HFWOA-VANET has the features of both meta-heuristic algorithms and is implemented to enhance the performance of VANET. This process is mainly based on consideration of Quality of Service (QoS) parameters of each vehicle. Therefore, the performance of vehicle can be determined and the better service in VANET platform is enabled. The implementation of this work is done on NS2 platform and the obtained results are analysed for ensuring the performance of the proposed model. Moreover, the performance of the model is compared with the existing technology; therefore, the proposed model can be ensured as a more effective technique than the existing technique in terms of performance metrics.
  Keywords: VANET; firefly optimisation algorithm; whale optimisation algorithm; QOS; QMM-VANET; HFWOA-VANET.
  
  
• Performance analysis of downlink precoding techniques in massive MIMO under perfect and imperfect channel state information in single and multi-cell scenarios  
  by Chanchal Soni, Namit Gupta 
  Abstract: The novel Optimised Max-Min Zero forcing precoder (OM2ZFP) scheme is proposed in this work. The optimization is incorporated with the chimp optimization strategy (CPO) to maximise the spectral efficiency, achievable sum rate, max-min rate, and minimise BER. The designed precoder model is contemplated under single cell perfect CSI, single-cell imperfect CSI and multiple cells perfect CSI, multi-cell imperfect CSI. Three pre-coding schemes, zero forcing (ZF), Maximum Ratio Pre-coding (MRT) and Minimum Mean Square Error (MMSE) precoder techniques, are implemented in the Matlab platform to manifest the effects of the novel designed precoder. The performance of the achievable sum rate is analysed under three cases, namely case I (fixed users and varying antenna), case II (fixed and varying) and case III (varying channel estimation error). The results show that the increasing number of antenna and users enhance the spectral efficiency, downlink transmits power and achievable sum rate performance.
  Keywords: massive MIMO; precoder; downlink transmission; antenna; optimisation; spectral efficiency; achievable sum rate.
  
  
• Preoperative staging of endometrial cancer based on decision tree model  
  by Jun Xu, Hao Zeng, Shuqian He, Lingling Qin, Zhengjie Deng 
  Abstract: Endometrial cancer is extremely common in gynaecological tumours. Ultrasound technology has become an important detection method for endometrial cancer, but the accuracy of ultrasound diagnosis is not high. Therefore, using data-driven methods to accurately predict the preoperative staging of endometrial cancer has important clinical significance. To build a more accurate diagnosis model, this paper uses a decision tree model to analyse the preoperative staging diagnosis indicators of endometrial cancer. Experimental results show that the three-detection data of tumour-free distance (TFD), ca125, and uterine to endometrial volume ratio are of high value for the diagnosis of endometrial cancer. The accuracy, sensitivity and specificity of the random forest (RF) model based on decision tree for preoperative staging of endometrial cancer were 97.71%, 94.11% and 100.00%, respectively. The comprehensive predictive ability based on the RF model has good application value for the prediction of preoperative staging of endometrial cancer.
  Keywords: random forest; decision tree; machine learning; endometrial cancer; preoperative staging.
  
  
• RPL-OFs analysis and dynamic OF selection for QoS optimisation of RPL protocol  
  by Sharwari Solapure, Harish Kenchannvar, Umakant Kulkarni 
  Abstract: Quality of Service (QoS) requirements differ for various IoT applications, such as smart health reliability is the need, for industry delay is essential etc. The Routing-Protocol-for-Low-Power-Lossy-Network (RPL) with Objective Function (OF) is used for routing in an IoT application. Default RPL-OF is deficient to fulfil the QoS requirements of different IoT applications. Hence, several OF designs were proposed as per the QoS need in the earlier research. The work presented in this paper is the extension of previous research work. The analysis of these OF designs is carried out with the parameters such as number of nodes, simulation-time, data-rate, Media Access Control (MAC) protocols, communications ranges and different topologies. This analysis resulted in a dataset that addresses most of the QoS-requirements and it is used to optimise the RPL protocol QoS performance. Decision tree algorithm is used to predict a suitable RPL-OF design. The accuracy achieved using Gini and Entropy method of decision tree is 87.14% and 88.57% respectively. Thus, the contribution of this research is to prepare the dataset using comprehensive analysis and use the same for predicting suitable RPL-OF design according to QoS-requirements of an IoT application. The proposed methodology is useful in IoT applications where dynamic-OF selection as per QoS requirements is needed.
  Keywords: IoT; RPL; OF; LLN; QoS.
  
  
• An improved fuzzy clustering log anomaly detection method  
  by Shuqian He, WenJuan Jiang, Zhengjie Deng, Xuechao Sun, Chun Shi 
  Abstract: Logs are semi-structured text data generated by log statements in software code. Owing to the relatively small amount of abnormal data in log data, there is a situation of data imbalance, which causes a large number of false negatives and false positives in most existing log anomaly detection methods. This paper proposes a fuzzy clustering anomaly detection model for unbalanced data, which can effectively deal with the problem of data imbalance and can effectively detect singular anomalies. We introduce an imbalance compensation factor to improve the fuzzy clustering method, and use this method to build an anomaly detection model for anomaly detection of real log data. Experiments on real data sets show that our proposed method can be effectively applied to log-based anomaly detection. Furthermore, the proposed log-based anomaly detection algorithms outperform other the state-of-the-art algorithms in terms of the accuracy, recall and F1 measurement.
  Keywords: distributed information system; log data; anomaly detection; artificial intelligence for IT operations; fuzzy clustering; imbalanced datasets; unsupervised learning; machine learning.
  
  
• Research on system parameter optimisation in electromagnetic tomography technology  
  by Liu Li, Yue Luo, Yao Huang, Lijuan Wu 
  Abstract: Electromagnetic tomography technology (EMT) based on the principle of electromagnetic induction is a multi-phase flow detection technology. It is reconstructed without contact and intervention. The development process of EMT is presented in this paper. The basic physical model is constructed. The internal sensitive field equation is given. The detection values are analysed by the numerical calculation method. It is mainly to establish the sensitivity model and the detection value matrix. By using the control variable method, the effects of the excitation current frequency, current strength on the detection values and phases are compared and analysed. Under the same parameter setting conditions, different imaging algorithms are used to reconstruct the images for models. In the inverse problem, Tikhonov regularization, LBP methods and conjugate gradient algorithm are introduced. The optimal parameters are determined by using parameters of IE (Image Error) and CC (Correlation Coefficient) to evaluate the reconstructed image.
  Keywords: electromagnetic tomography technology; image reconstruction; Ccnjugate gradient algorithm; inverse problem.
  
  
• OLSR-ETX: a parameterised solution for oscillatory network packet losses  
  by Kifayat Ullah, Ihtisham Ali 
  Abstract: Expected Transmission Count (ETX) has gained popularity due to identifying a high-throughput path in the multihop wireless network. However, the oscillatory network may not work correctly with a high traffic load; the probe packets may be lost or queued. This paper proposes a parameterized solution (data rate tuning and packet size adjustment) to minimize packet losses. Experimental results indicate that the network's performance has improved using ETX as a routing metric by tuning data rates and adjusting packet size. The results show that by keeping the Data rate under 200kbps and a Packet size of 256 bytes, the performance of the OLSR-ETX routing protocol has improved in the oscillatory network. Finally, we have evaluated the OLSR-ETX parameterized-based solution with OLSR-ETX in oscillation scenarios concerning packet loss ratio. The results show that a parameterized-based solution improves the functionality of the routing protocols in the oscillatory network.
  Keywords: ETX; OLSR-ETX; OLSR; oscillatory network; packet loss ratio.
  
  
• An efficient blockchain model for improving data transmission rate in ad hoc networks  
  by Lucky Narayana 
  Abstract: A Mobile Ad hoc Network (MANET) is an infrastructure-less network that can be established dynamically whenever and wherever required for establishing communication. The MANET is a series of nodes with capabilities in wireless communication and networking. A temporary network that is possible without an already-oriented network or centralised supervisor is linked by an ad hoc network to its mobile hosts as required. The topology of an ad hoc network is different for node mobility. The function of the ad hoc network needs its own solutions and should be different from the static networks to build applications. Radio nodes are immediately established to communicate with each other. With the help of intermediate nodes, nodes not within each other\'s radio range can be transmitted from source to destination. As ad hoc networks are dynamic in nature, they frequently undergo several attacks that reduces the data transmission rate. In the proposed work, an efficient blockchain model is used in ad hoc networks for improving the data transmission rate by analysing the cause for packet loss. In the proposed model, a Malicious Task Identification Head Node (MTIHN) is selected from the network that analyse the blocks generated after every transaction for checking the cause of packet drops. The blockchain is a modern data storage platform. In the various systems with different operating principles this does not operate in the same way. The proposed work explores network security using the blockchain framework to make it easier to send messages and information without loss that improves system performance. The proposed model is compared with the traditional methods and the results show that the proposed model exhibits better performance in improving Data Transmission Rate.
  Keywords: data transmission rate; malicious actions; blockchain; security; ad hoc networks; block generation.
  
  
• Research on wireless routing problem based on dynamic polycephalus algorithm  
  by Zhang Yi, Yang Zhengquan 
  Abstract: The efficiency of the traditional Physarum Polycephalum Model (PPM) is low for wireless planning problems. Also, other heuristic algorithms are easy to fall into local optimum and usually require a large training set to find the optimal parameter combination. Aiming at these problems, we propose a new dynamic model of Physarum Polydynia (DMOP2) algorithm combined with PPM in this paper. This algorithm can judge the irrelevant nodes according to the traffic matrix after each iteration and then delete them and re-establish a new distance matrix when solving the routing problem. The improvements not only reduce the time consumed by calculation but also improve the accuracy of calculation pressure. Simulation experiments in random network and real road network prove the feasibility and effectiveness of the proposed algorithm in solving the path planning problem, and the experimental results show that the efficiency is significantly improved compared with PPM.
  Keywords: wireless planning; Physarum Polycephalum model; dynamic model.
  
  
• Energy-efficient dynamic load balanced clustering for MANET  
  by Naghma Khatoon, Vinay Singh, Prakash Kumar 
  Abstract: In mobile ad-hoc network (MANET), enhancing network lifetime is a challenging issue. Clustering is proved to be a suitable solution to increase scalability and MANET lifetime. In this paper, we present an energy-efficient dynamic load-balanced clustering for MANET. For cluster formation, nodes are divided into open set and restricted set. Depending upon the weight of Cluster Head (CH), node join them to form a cluster which make it load balanced. We use technique for self-adjustment of role of CHs dynamically based on fitness factor which is derived from remaining energy and weight of nodes to increase CH lifetime. The proposed method is experimented extensively and compared with related existing algorithms to demonstrate its ascendancy related to various performance metrics like packet delivery ratio, network lifetime, average number of clusters formed and re-clustering required. Also, we demonstrate that the work proposed accomplishes persistent messages and the time complexity is linear.
  Keywords: MANET; cluster head; fitness factor; remaining battery energy; packet delivery ratio.
  
  
• Research on facial expression recognition based on multimodal data fusion and neural network  
  by Yi Han, Xubin Wang, Zhengyu Lu 
  Abstract: Facial expression recognition is a challenging task when a neural network is applied to pattern recognition. Most of the current recognition research is based on single source facial data, which generally has the disadvantages of low accuracy and low robustness. In this paper, a neural network algorithm of facial expression recognition based on multimodal data fusion is proposed. The algorithm is based on the multimodal data, and it takes the facial image, the histogram of oriented gradient of the image and the facial landmarks as the input, and establishes a convolutional neural network designed to extract features from facial image, a neural network designed to extract features from facial landmarks, and a neural network designed to extract features from histogram of gradient, and three sub-neural networks to extract data features, using multimodal data feature fusion mechanism to improve the accuracy of facial expression recognition. Experiment results show that, the algorithm has a great improvement in accuracy, robustness and detection speed.
  Keywords: multimodal data; deep learning; neural network; facial expression recognition; data fusion.
  
  
• A hybrid malicious node detection approach based on fuzzy trust model and Bayesian belief in wireless sensor networks  
  by Wuchao Shi 
  Abstract: With the wide range of wireless sensor network (WSN) applications including environmental monitoring and healthcare, the sensor nodes in WSN are susceptible to security threats including dishonest recommendation attacks from malicious nodes, which could disrupt communications integrity. Thus, malicious node detection in WSN is essential. In recent years, several malicious node detection approaches based on trust management were proposed to protect the WSN against dishonest recommendation attacks. However, the existing approaches ignore data consistency and re-evaluation of participating nodes in trust evaluation, which seriously undermine their effectiveness. To address these limitations, we propose a hybrid malicious node detection technique for WSN based on the fuzzy trust model (FTM) algorithm and the Bayesian belief estimation (BBE) approach. The key idea in the proposed approach is to determine direct trust values through the FTM algorithm using the correlation of data collected over time and ascertain the trustworthiness of indirect trust values from recommendation nodes via the BBE approach.
  Keywords: wireless sensor network; dishonest recommendation attacks; fuzzy trust model; Bayesian belief.
  
  
• A trusted management mechanism based on trust domain in hierarchical internet of things  
  by Mingchun Wang, Jia Lou, Yedong Yuan, Chunzi Chen 
  Abstract: Existing trusted models usually authenticate the identity and behaviour of sensing nodes, without considering the role of sensing nodes in the process of interaction and transmission of information. Therefore, in view of the hierarchical wireless sensor network architecture of the internet of things, this paper proposes a new hierarchical trusted management mechanism based on trusted domain. The mechanism abstracts different nodes in the hierarchical structure of the internet of things, gives them different identities, and calculates the trust value of the sensing nodes by using similarity weighted reconciliation method. The experimental results show that the proposed scheme is feasible and effective.
  Keywords: trusted domain; trusted management; similarity weighted reconciliation; trust value; hierarchical structure.
  
  
• Task scheduling in multi-cloud environment via improved optimisation theory  
  by Prashant Balkrishna Jawade, S. Ramachandram 
  Abstract: One of the most popular technologies nowadays, cloud computing has a big demand in the distributed software space. It is highly difficult for CSPs to work together in a multi-cloud context, and contemporary literature does not adequately address this issue. Throughout this work, a protected TS paradigm in a multi-cloud environment is introduced. The suggested scheme mainly focuses on the optimal scheduling of tasks by considering a Modified Deep Neural Network (DNN) as a task scheduler. Accordingly, the task is allotted based upon makespan, execution time, security constraints (risk assessment), utilisation cost, maximal service level agreement adherence, and power usage effectiveness. Moreover, the weights of DNN are tuned optimally by self-improved aquila optimisation technique. The developed model has a lowMAE value of 0.052581, which is 46.67%, 90.85%, 89.29%, and 86.43% better than DNN, NN, RNN, and LSTM, respectively.
  Keywords: task scheduling; execution time; modified DNN; risk assessment; SI-AO model.
  
  
• A new time-frequency synchronisation algorithm based on preamble sequence in OFDM system  
  by Weimin Hou, Yan Wang, Yanli Hou 
  Abstract: Aiming at the problems of high computational complexity in the timing synchronization phase and poor frequency offset estimation performance of existing time-frequency synchronization algorithms, this paper proposed an improved time-frequency synchronization algorithm based on preamble sequence for OFDM systems. The preamble sequence is designed by using the property that the cross-correlation value of the Constant Amplitude Zero Auto Correlation (CAZAC) sequence with different root values is close to zero. Based on its features, a timing metric function and the frequency offset estimation function are designed. The frequency offset estimation function is used to obtain the coarse fractional frequency offset, and the fine fractional frequency offset is obtained by combining cyclic prefix (CP) and cyclic suffix (CS). Then the time domain sliding correlation between receiving sequence and the local preamble sequence is used to estimate the integer frequency offset. The results indicate that the proposed method has better synchronization capability than existing algorithms.
  Keywords: OFDM system; timing synchronization; frequency offset estimation; preamble sequence; CAZAC sequence.
  
  
• Automatic modulation recognition based on channel and spatial attention mechanism  
  by Tianjun Peng, Guangxue Yue 
  Abstract: With the complexity of the wireless communication environment, automatic modulation recognition (AMR) of wireless communication signals has become a significant challenge. Most existing researches improve the model recognition performance by designing high-complexity architectures or providing supplementary feature information. This paper proposes a novel AMR framework named CCSGNet. The convolutional neural network (CNN) and bidirectional gate recurrent unit (BiGRU) are employed in CCSGNet to reduce the spectral and time variation of the signals, furthermore, the channel and spatial attention are employed to fully extract local and global features of signals. In order to reduce the training time cost of the model, we propose a piecewise adaptive learning rate tuning method to improve the training of the model. The comparisons with several common learning rate tuning methods on CCSGNet show that the proposed method achieves convergence in 25 training epochs, reducing the training time cost of the model. Moreover, CCSGNet improves the recognition accuracy of 16QAM and 64QAM by 6.47%-50.95% and 4.54%-25.66%, respectively.
  Keywords: automatic modulation recognition; attention mechanism; learning rate; deep learning.
  
  
• Optimisation of a high-speed optical OFDM system for indoor atmospheric conditions  
  by B. Sridhar, S. Sridhar, Naresh K. Darimireddy 
  Abstract: VLC provides high security and broadband functionality for optical communication in free space. In particular, this proposed work focuses on analysing receiving power distribution patterns and signal-to-noise ratios for indoor and vehicle applications. The optical systems of indoor communications are more suitable than wireless radio systems. The significant advantage of optical wireless communication (OWC) is providing high-speed data up to 2.5 Gbps at a low cost. In indoor areas such as auditoriums and public places, the OWC systems are more suitable. But optical signals are distorted by the signal propagation effects due to obstacles, walls, etc. The proposed system is an OFDM-based system that can transmit multiple channels and connects many modems over a given indoor area. Proposed methods initially focus on the LED/LD transmitter sources placement at the ceiling of indoor space and observed signal power distribution; in an IM/DD-based OWC system, the information signal must be accurate and nonnegative. The proposed asymmetric optical OFDM (ACO-OFDM) system is implemented for indoor communications, and the system's performance is evaluated with the Bit error rate. In particular, the performance of the specific M-QAM ACO-OFDM method with adaptive frequency is assessed by using theoretical analysis and simulations. Compared to the M-QAM ACO-OFDM method, the ACO-OFDM and DCO-OFDM showed lower spectral efficiency performance for the OWC system in the frequency selective channel.
  Keywords: ACO-OFDM; indoor networks; power distribution; clipping; bit error rate.
  
  
• A prototype model of an unmanned automated railway level crossing traffic control system using ultrasonic and infrared sensors  
  by Priyadarshi Guha, Sukanya Kool, Rishabh Pipalwa, Aakash Acharjee, Abhijit Paul 
  Abstract: The accessibility of the railway has expanded as the population has grown, and controlling the level crossing in an unmanned automated way is necessary to ensure the safety of the public. This paper proposes an automatic railway level crossing control system which not only made level crossing automated but also introduces checking of the presence of any obstacles when the railway gate is closed at the time when a train is approaching. An ultrasonic sensor is used to monitor the presence of any obstacle in the restricted area in between the railway gates only when railway gates are closed. An infrared sensor is also used only to check whether the train has entered or passed away from that restricted area of the railway gate zone. The infrared sensor and ultrasonic sensor used in the proposed model have achieved accuracy up to 99.02% and 99.06%, respectively.
  Keywords: level crossing; ultrasonic sensor; infrared sensor; railroad; train.
  DOI: 10.1504/IJWMC.2023.10062691
   
  
• Performance evaluation and direction of arrival estimation in diverse beamforming approaches for enhanced signal reception in wireless communication systems  
  by Zareena Amin, Naresh Kumar, Subhash Dubey 
  Abstract: The 5G technology has revolutionised the tech, delivering gigabits per second speed. To fulfill the data demands of end-users seamlessly, advanced signal processing techniques have been implemented like, zero forcing, diversity implementation, beamforming etc. This article offers a comparative assessment of beamforming algorithms, including Minimum Variance Distortionless Response (MVDR), Beamscan, Multiple Signal Classification (MuSiC), and Linearly Constrained Minimum Variance (LCMV). The evaluation revolves around the Direction of Arrival (DoA) estimation for incoming signals, conducted under realistic conditions. Among the algorithms under scrutiny, the MuSiC algorithm exhibits superior performance. Therefore, we extend the scope of the discussion by including an in-depth analysis on the error performance of the MuSiC algorithm. Different factors that can impact the accuracy of the estimating procedure have been considered. Results depict that Signal to Noise ratio, Snapshot count and Array size positively correlate with estimation accuracy. Also, for coherent signals an enhanced MuSiC algorithm is presented.
  Keywords: array signal processing; beamforming algorithms; DoA estimation; mean square error; MuSiC algorithm.
  DOI: 10.1504/IJWMC.2024.10062853
   
  
• An end-to-end instance segmentation method based on improved ConvNeXt V2  
  by Wenlu Wang, Sun Yin, Manman Xu, Dongxu Bai, Li Huang, Chunlong Zou, Baojia Chen, Dalai Tang 
  Abstract: The detecting work of mobile robots may be hampered by indoor clutter, moving people, and uneven lighting, which will also impair the robots' subsequent feedback to the environment. This study first enhances the performance of the existing RTMDet model by using the more potent ConvNeXt V2 as the model's backbone network. It then uses the adaptive search NAS-FPN as the fusion network to realize network detection at any time. Finally, it selects AdamW as the model's optimizer, which resolves the issue of excessive memory usage when updating parameters and enhances the model's performance. and improves the performance of the model. The improved model utilized in this study performed well in the experiments, and after training and testing on the Cityscapes dataset, the recognition accuracy was 65.0%.
  Keywords: instance segmentation; ConvNeXt V2; deep learning; optimiser; feature fusion.
  DOI: 10.1504/IJWMC.2024.10063034
   
  
• Multi-stream fusion network for continuous gesture recognition based on sEMG  
  by Jun Li, Chunlong Zou, Dalai Tang, Sun Yin, Hanwen Fan, Li Boao, Xinjie Tang 
  Abstract: Dynamic gesture recognition is an essential step in human-computer interaction. However, dynamic gesture recognition based on surface electromyography (sEMG) signals faces issues such as incomplete feature extraction and incomplete recognition of temporal information between continuous gestures. Therefore, this paper proposes a multi-stream fusion network (MSK-LCNN) for dynamic gesture recognition to improve accuracy. We combine CNN and LSTM models into a unified framework which extracts spatio-temporal information both globally and in depth, and combines feature fusion to retain essential information. This framework uses an attention mechanism (SKNet) to learn more intricate feature information. The recognition accuracy of this method on our dataset is 95.23%, and on the NinaPro DB1 dataset, it is 91.45%, which outperforms other similar networks in recent years. Applying this algorithm to prosthetic hand control has achieved flexible and stable control of the prosthetic hand.
  Keywords: sEMG signals; gesture recognition; attention mechanisms; neural networks; multi-stream network.
  DOI: 10.1504/IJWMC.2024.10063096
   
  
• Throughput and BER analysis in LoRa technology for performance enhancements using dynamic measurement  
  by Prajakta Amol More, Zuber M. Patel 
  Abstract: We have observed that the LoRa network efficiency is influenced by several factors like throughput and Bit Error Rate (BER). Considering these in wireless network long range with better reliability and throughput is the demand for ongoing applications. LoRa technology shows promising results with its LPWAN protocol in IoT applications. Throughput plays a major role which we will be looking into and focusing to enhance its output. Bit Error Rate has a significant impact over LoRa communication network. Bit Error Rate with SNR need to be balanced with longer range, greater throughput. So these factors are optimised to improve the performance of LoRa network. Our paper presents throughput and BER with the changes in physical layer parameters of the LoRaWAN technologies we can design a dynamic parameter change algorithm for it and which can improve the overall performance of the system. Results are justified using MATLAB simulations.
  Keywords: LPWAN; physical layer; throughput; bit error rate.
  DOI: 10.1504/IJWMC.2024.10063169
   
  
• Perception and early warning technology of power substation safety construction operation based on multi-dimensional information fusion  
  by Yan Ke, Hongtao Chen, Zhiyuan Liu, Zhiyong Yang 
  Abstract: In response to the increasing requirements of on-site construction or maintenance work in power substations, leading to frequent safety control issues, this paper proposes a power substation safety construction operation perception and early warning technology based on multidimensional information fusion. First, based on the safety construction requirements within the substation, such as interlocking, equipment retrofitting, and defect elimination, the paper designs the layout of sensors for electromagnetic intensity, meteorology, toxic gases, oxygen concentration, as well as monitoring devices like video images, to enhance the perception capability of power construction operations within the substation. Secondly, the paper studies the cloud-edge collaborative multidimensional information transmission and fusion technology to achieve comprehensive management and application of on-site data within the substation. Finally, key technologies such as transfer learning, semantic segmentation, anomaly detection, and real-time data analysis are utilized to achieve real-time early warning for power substation construction operations.
  Keywords: power substation; multi-dimensional information; safety construction operation; sensor.
  DOI: 10.1504/IJWMC.2024.10063196
   
  
• Research on fault diagnosis for power transformer based on random forests and wavelet transform  
  by Ming Zhang, Chongfeng Fang, Shuang Ji 
  Abstract: In order to improve the accuracy of power transformer fault diagnosis and condition monitoring, this paper proposes a fault diagnosis method for power transformers based on Random forests (RFs) and wavelet transform. Firstly, the wavelet transform method is adopted to decompose the noisy vibration signal into multi-scales, and then the detailed signals at different scales are processed to achieve fault feature extraction of the power transformer vibration signals. Secondly, the mapping relationship between fault features and fault types of vibration signals is established by RFs algorithm, and the fault diagnosis model is trained by RFs algorithm. Finally, by identifying the experimental data of the normal and fault states of the power transformers, the accuracy reached 96.52%, which is suitable for monitoring and diagnosing the different working states of the power transformers.
  Keywords: fault diagnosis; power transformer; random forests; wavelet transform.
  DOI: 10.1504/IJWMC.2024.10063197
   
  
• Multi-task learning neural network for monitoring and diagnosis of smart meters in power IoT systems  
  by Ming Zhang, Yong Cui, Lei Wang, Shuang Ji 
  Abstract: With the continuous development of power Internet of Things technology, the scale and complexity of power systems are increasing, and the safe operation of power IoT systems is facing challenges In this paper, an intelligent monitoring and diagnosis model is proposed for smart meter, a key terminal in an IoT system The proposed model is developed using deep learning methods by proposing a novel multi-task learning deep neural network (DNN), which is designed for online perception, monitoring and diagnosis of the operation status of smart power electric meters The proposed method can simultaneously deal with both the regression task for meter pointer localization and the classification task for safety status identification, which support to realize multi-task-oriented deep data mining and intelligent decision-making The experimental results of this paper verify the accuracy and reliability of the proposed method for a real power IoT system, and validates the effectiveness of the method
  Keywords: deep neural network; multi-task learning; computer vision; smart meter; power IoT terminal.
  DOI: 10.1504/IJWMC.2024.10063330
   
  
• Research on three-dimensional perception and protection technology for power construction safety operations  
  by Yan Ke, Hongtao Chen, Zhiyuan Liu, Zhiyong Yang, Lin Song 
  Abstract: The traditional safety supervision of power transmission and transformation operation sites mainly relies on manual and video surveillance, resulting in a relatively low level of intelligent safety supervision in power construction operations This paper aims to conduct research on key technologies for three-dimensional perception and protection of power construction operations to enhance intelligent supervision in China's power construction operations Firstly, it reviews the research progress of power construction operations at home and abroad to provide references for intelligent supervision of power construction operations Secondly, based on the characteristics and safety requirements of power transmission and transformation operations, it conducts research on four key technologies, including three-dimensional protection based on lidar and image recognition, intelligent monitoring of environmental and hazardous gases based on various sensing devices, high-voltage near-electric detection based on digital filtering, and the development of a platform for three-dimensional protection safety warning monitoring of power construction These studies aim to improve the level of intelligent safety management on power construction. Lastly, this paper discusses further research directions including multi-sensor fusion, data analysis and processing, real-time warning and decision support, and virtual reality technology to achieve comprehensive monitoring and intelligent management of power construction, thereby improving safety and efficiency.
  Keywords: power construction operation; three-dimensional perception and protection; image recognition; near-electric detection; sensors; platform.
  DOI: 10.1504/IJWMC.2024.10063331
   
  
• Research on detection technology of appearance intellectual property based on convolutional neural network  
  by Haiyan Huang 
  Abstract: Aiming at the problem of low accuracy of traditional cigarette box appearance intellectual property detection, an appearance intellectual property detection method based on CNN-LSTM is proposed. CNN network and LSTM network are adopted to extract the spatial and temporal features of cigarette boxes respectively, and the fusion processing is carried out. Then, the multi-branch parallel appearance segmentation model is used to segment images. The results show that the detection accuracy of the proposed model is 99.81%, which is 4.21%, 8.98% and 10.96% higher than that of Inception, ResNet and MobileNet networks, respectively. This shows that the method can improve the accuracy of intellectual property detection of cigarette boxes.
  Keywords: convolutional neural network; LSTM; cigarette boxes; intellectual property; appearance detection.
  DOI: 10.1504/IJWMC.2024.10063372
   
  
• A method of selecting characteristics based on P_KPCP for new power system operation mode  
  by Xiaoli Guo, Qingyu Shan, Zhenming Zhang, Hao Jiang 
  Abstract: The characteristic variables of the new power system operation mode are developing in the direction of high dimensionality and diversification, which leads to the lack of applicability and accuracy of the traditional characteristic variable screening methods. For this reason, a P_KPCP-based feature selection method for new operating modes of power systems is proposed. First, in order to quantify the correlation among various operating variables, a correlation quantification method based on the Pearson coefficient of the characteristic variable of the operating mode is designed. Then, reduce the dimension of feature vectors, reduce the impact on the accuracy of operation mode extraction, and build a P_KPCA-based screening model for strongly correlated feature variables of operation mode. Finally, based on more than 8000 power grid operation sections, experiments are carried out to verify the accuracy and rationality of the method in this paper.
  Keywords: power operation data; new power system; feature dimension reduction; kernel principal component analysis; dimension reduction.
  DOI: 10.1504/IJWMC.2023.10063373
   
  
• Deep learning-based wall crack detection  
  by Zujia Zheng, Kui Yang 
  Abstract: Aiming at the problems of large weight and high complexity of YOLOv4 network, this paper proposed an improved wall crack detection method based on YOLOv4 network. Firstly, the backbone feature extraction network of YOLOv4 was replaced by Mobile Netv2, and then the common convolution in deep network was replaced by deep separable product, so as to make the whole network model lightweight. Then, SENet attention mechanism is incorporated to make up for the loss of detection accuracy caused by network lightweight. Finally, the data set was constructed and annotated by collecting crack images and data amplification. The experimental results show THAT this method can greatly reduce the network weight, the number of parameters and the computation, and shorten the detection time while maintaining a high detection level, which can meet the needs of various wall crack detection work.
  Keywords: YOLOv4; wall crack detection; target detection.
  DOI: 10.1504/IJWMC.2024.10063608
   
  
• Performance evaluation of centralised and distributed controllers in software defined networks  
  by Houda Hassen, Soumaya Meherzi 
  Abstract: Software-Defined Networking (SDN) is a new network paradigm that has been proposed to overcome the limitations of traditional networks in terms of complexity and management. According to SDN architecture, the control plane and the data plane are decoupled, and all the network’s intelligence is centralized in the controller. A myriad of open-source controllers have been proposed for both research and industrial use. Such controllers, which are designed with different characteristics, have been shown to behave differently with regard to some network QoS parameters. Due to the crucial role of the controller in SDN networks, we intend to provide in this paper a thorough performance analysis of the most commonly used controllers. As centralized controllers, we have considered POX and RYU, whereas OpenDayLight (ODL) and ONOS have been chosen as examples of distributed controllers. For each kind of controller, the performance behavior is analyzed under different topology types and various network scales using a wide range of QoS metrics.
  Keywords: SDN controller; POX; RYU; OpenDayLight; ONOS; performance parameters.
  DOI: 10.1504/IJWMC.2023.10063687
   
  
• DBSCAN behaviour analysis and N-Adaboost prediction model research for mathematics majors academic prediction  
  by Xiaoni Zhang 
  Abstract: From the perspective of distance optimisation, a density-based spatial clustering applied noise algorithm is proposed for clustering analysis and academic prediction of mathematical students' behaviour. This algorithm improves the clustering effect and accuracy by improving the selection of neighborhood radius. Secondly, to address the limited learning performance of a single classifier, an N-Adaboost model based on multiple classifiers is proposed. The experiment shows that when the number of clusters is 4, the network behavior description index reaches the optimal level, with a maximum contour coefficient of 0.667. The N-Adaboost prediction model has high accuracy, accuracy, and recall rate. When N = 3, the model has the best performance and can successfully predict and analyse data. In summary, the density based noisy clustering algorithm based on distance optimisation and the N-Adaboost prediction model based on multiple classifiers have broad application prospects in student behaviour clustering analysis and academic prediction problems.
  Keywords: mathematics major; academic prediction; DBSCAN; N-Adaboost; distance optimisation.
  DOI: 10.1504/IJWMC.2024.10063765
   
  
• SSD object detection algorithm based on knowledge map  
  by L. Huang, Xiaofeng Wang, Jianhua Lu, Wei Hu, Changrong Zhang 
  Abstract: With the application of artificial intelligence to all aspects of people's lives, talent is the first resource. Through the analysis of the current situation of talent training in colleges and universities, it is found that the major has problems such as scattered knowledge points, overlapping content, single practice path, and lack of effective evaluation. In view of the above problems, this paper proposes a multidisciplinary and comprehensive practical teaching mode based on knowledge graph by conducting research on the indoor object detection algorithm of service robot by SSD algorithm, and discusses multi-path practical teaching and evaluation methods from the aspects of teaching objectives, problem decomposition, resource integration, implementation methods and grade evaluation. Promote the all-round development of students by learning the knowledge of object detection.
  Keywords: knowledge map; integrated practice; SSD object detection algorithm.
  DOI: 10.1504/IJWMC.2024.10063828
   
  
• Low-complexity detector performance evaluation for cell-free massive MIMO systems  
  by Mitesh Solanki, Shilpi Gupta 
  Abstract: In 5G, massive MIMO combines beamforming and spatial multiplexing of users to enhance spectral efficiency and energy efficiency It employs a centralized solution to densify a network, resulting in intercellular interference because of its cell-centric design Massive MIMO systems enable user-centric transmission in cellular networks to overcome inter-cell interference limitations and provide macro-diversity Massive MIMO distinguishes itself from coordinated, distributed wireless systems with its scalability and cell-free design This article investigates the immense potential of our proposed detector, this promising technology, and practical deployment issues due to signal coprocessing's increased backhauling overhead We propose a conjugate gradient (CG)-based detector that enables faster convergence and smaller approximated errors for cell-free massive MIMO systems An analytical approach describes the asymptotic performance of a conjugate gradient-based likelihood ascent search detector with massive antennas Additionally, the joint channel estimation and data detection (JCD) method is used in real-world scenarios with imperfect channel state information
  Keywords: massive MIMO; conjugate gradients; cell-free massive MIMO; joint channel estimation and data detection; imperfect channel state information.
  DOI: 10.1504/IJWMC.2024.10063829