Current Issue

30 December 2021, Volume 5 Issue 4

Previous Issue    Next Issue

Frontier and Comprehensive Review

Graph neural network driven traffic prediction technology:review and challenge

Yi ZHOU, Shuting HU, Wei LI, Nan CHENG, Ning LU, Xuemin(Sherman) SHEN

2021, 5(4):  1-16.  doi:10.11959/j.issn.2096-3750.2021.00235

Asbtract ( 1561 )   HTML ( 448)   PDF (1848KB) ( 1446 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

With the rapid development of Internet of things and artificial intelligence technology, accurate analysis and prediction of traffic data have become the primary target of intelligent transportations.In recent years, the method of traffic forecasting has gradually changed from the classical model-driven type to the data-driven type.However, how to effectively analyze the spatial-temporal characteristics of road networks through big data is one of the key issues in the traffic prediction process.Spatiotemporal big data analysis is a powerful tool for the traffic prediction.The traffic network can be modeled as a graph network, while the deep learning method can be extended on the graph network.Utilizing graph neural networks, we can build the spatiotemporal prediction model, and obtain the spatial-temporal correlation between the sensor nodes in road networks effectively by using graph convolution, which can significantly improve the accuracy of traffic prediction models.The traffic forecasting technology driven by graph neural networks was explored, and two kinds of traffic prediction models based on the analysis of deep spatial-temporal characteristics were extracted.The actual cases were analyzed and evaluated to discuss the technical advantages and key challenges of graph neural networks in the traffic prediction.The potential issues of graph neural network driven prediction mechanisms were also excavated.

Theory and Technology

Research on LoRa network security schemes based on RF fingerprint

Yu JIANG, Siqing CHEN, Wen SUN

2021, 5(4):  17-25.  doi:10.11959/j.issn.2096-3750.2021.00228

Asbtract ( 393 )   HTML ( 63)   PDF (2542KB) ( 462 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Long range radio (LoRa) is widely used in the IoT due to its advantages of long distance and low power consumption.However, LoRa network has no reliable security scheme currently, making it unable to guarantee the communication security.Therefore, based on the uniqueness and tamper-resistance of radio frequency fingerprint, it was proposed to receive radio frequency signals of the LoRa end nodes which requested access, extract the fingerprints from them, mark it and match with the customized multi-scale security rules according to demands to decide whether the identities of the LoRa end nodes were safe, taking security measures accordingly.Based on this, original LoRa gateway and LoRa network architecture were improved, new workflows were designed, and two LoRa network security schemes were proposed.The two LoRa network security schemes were proposed which implement identity authentication and access control of the LoRa end nodes from the physical layer.It is only needed to improve the original LoRa gateway in the LoRa network architecture and its workflow, which adds new security measures and guarantees for LoRa applications on the basis of not affecting the original LoRaWAN security mechanism, with no need to modify a huge number of LoRa end nodes.The security schemes proposed have high practical value.

Research on location privacy protection of mobile terminals for maritime monitoring sensor networks

Xin SU, Su JIANG, Yiqing ZHOU

2021, 5(4):  26-36.  doi:10.11959/j.issn.2096-3750.2021.00244

Asbtract ( 205 )   HTML ( 40)   PDF (2988KB) ( 390 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Mobile edge computing can support various maritime applications with high reliability and low delay.However, many security problems in computing task offloading exist.The risk of location privacy leakage of maritime mobile terminals during task offloading was analyzed and quantified.The related location privacy protection model was established and a dynamic cache and spatial cloaking-based location privacy protection (DS-LPP) algorithm was proposed.Simulation results show that DS-LPP algorithm has better performance of constructing anonymous space and selecting relay node than traditional algorithms while protecting the location privacy of maritime mobile terminals.Therefore, the DS-LPP algorithm can be effectively applied to maritime monitoring sensor network with relatively scanty communication and computing resources, and ensure the continuity of location privacy protection.

FPGA-Jetson based hardware real-time co-simulation for smart grid

Tong DUAN, Venkata Dinavahi, Tianshi CHENG

2021, 5(4):  37-45.  doi:10.11959/j.issn.2096-3750.2021.00246

Asbtract ( 301 )   HTML ( 39)   PDF (8751KB) ( 242 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Smart grid is a power-communication coupled cyber-physical power system (CPPS), where the two-domain coupling features make the real-time simulation difficult.Different from the existing software-based co-simulation methods, a FPGA-Jetson based real-time co-simulation platform for smart grid by leveraging the FPGA’'s programmable computation capability and Jetson’s real-time operating system was designed.The power system simulation was carried out on the FPGA board, the information communication system simulation was completed on Jetson platform, and the two-domain information interaction was simulated through the PCIe channel.By making full use of the computing and data transmission capabilities of each module, the real-time simulation of the combination of power grid and communication network was realized.Finally, the corresponding specific modeling and parameters were designed for the two representative smart grid scenarios: hybrid AC-DC transmission network and micro-grid, the validity and scalability of the proposed real-time co-simulation architecture were verified.

Node selection based on label quantity information in federated learning

Jiahua MA, Xinghua SUN, Wenchao XIA, Xijun WANG, Hongzhou TAN, Hongbo ZHU

2021, 5(4):  46-53.  doi:10.11959/j.issn.2096-3750.2021.00249

Asbtract ( 382 )   HTML ( 63)   PDF (2528KB) ( 454 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Aiming at the problem that the difference of node data distribution has adverse effect on the performance of federated learning algorithm, a node selection algorithm based on label quantity information was proposed.An optimization objective based on the label quantity information of nodes was designed, considering the optimization problem of selecting the nodes with balanced label distribution under a certain time consumption limit.According to the correlation between the aggregated label distribution of selected nodes and the convergence of the global model, the upper bound of the weight divergence of the global model was reduced to improve the convergence stability of the algorithm.Simulation results shows that the new algorithm had higher convergence efficiency than the existing node selection algorithm.

Comparison of MIMO based on high capacity LPWAN technology TurMass^TM and LoRa

Hao JIANG, Hongming CHEN, Yilong CAO, Haoyang CUI

2021, 5(4):  54-61.  doi:10.11959/j.issn.2096-3750.2021.00243

Asbtract ( 373 )   HTML ( 50)   PDF (1789KB) ( 417 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

With the rapid development of low power wide area network (LPWAN) technologies, more and more IoT terminals need to access the network.In order to deal with the challenge of massive random access, grant-free random access with massive MIMO (MGFRA) was proposed for LPWAN scenarios.The MGFRA was introduced and the performance of MGFRA to validate its performance advantage of greatly improving the channel access capacity with extreme low signaling overhead was theoretically analyzed.Further, TurMass^TMwas introduced, which is a high-capacity LPWAN technology based on MGFRA.To demonstrate the advantages of TurMass^TMin practical applications, TurMass^TMand LoRa in signal coverage, communication rate, network capacity, low power consumption and low cost in typical LPWAN communication scenarios were compared.The comparison results show that TurMass^TMhas outstanding advantages in signal coverage, communication rate, network capacity, low power consumption and low cost compared with LoRa.

Research on DV-Hop location algorithm based on range correction and improved gray wolf optimizer

Xiaoqiang ZHAO, Shuai WU, Chuanyi GAO, Ning LI, Bodong LI, Xiaoyong YANG

2021, 5(4):  62-70.  doi:10.11959/j.issn.2096-3750.2021.00222

Asbtract ( 168 )   HTML ( 23)   PDF (1758KB) ( 414 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Node location is an important problem in wireless sensor network.Although the location algorithm based on distance measurement has small positioning error, it has many limitations when applied to outdoor environments.Therefore, based on the original distance vector-hop (DV-Hop) algorithm, received signal strength indication (RSSI) technology and the minimum mean square error (MMSE) criterion to modify the algorithm’s ranging process were introduced, and the improved gray wolf optimizer was used to optimize the process of determining the coordinates of unknown nodes.Simulation results show that, compared with the original DV-Hop algorithm and IPDV-Hop algorithm, the average location error rate of the IGDV-Hop algorithm under the initial parameters was reduced by 28% and 17% respectively, and the location effect was significantly improved.

Cognitive waveform design for radar-communication transceiver networks

Yu YAO, Yanjie LI, Lenan WU, Pu MIAO, Xiaoyu TANG

2021, 5(4):  71-80.  doi:10.11959/j.issn.2096-3750.2021.00232

Asbtract ( 152 )   HTML ( 20)   PDF (1584KB) ( 238 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

The system architecture for cognitive radar-communication (CRC) transceiver was proposed.A cognitive waveforms design approach, which is suitable for simultaneously performing both data communication and target detection was presented.This approach aims at estimating target scattering coefficient (TSC) from the radar scene and facilitating high-data-rate communications.In order to minimize the mean square error (MSE) of the TSC, a convex cost function was established.The peak to average power ratio (PAPR)-constrained optimal solution was achieved by applying the Kalman filtering-based strategy to design the set of ultra-wideband (UWB) transmission pulses and embed into them the information data with the M-ary position phase shift keying modulation technique.In addition to theoretical considerations, the simulation results show an improvement in TSC estimation and target detection probability as the number of iterations increases, while still transmitting data rates in the range of several Mbit/s with low bit error rates between CRC transceivers.

Neural network wind speed prediction based on multiple prediction model and nonlinear combination

Jiajun WANG, Wei CAO, Guilong ZHANG, Huaizhi ZHANG, Zixing LING, Xiaoqiang ZHAO

2021, 5(4):  81-89.  doi:10.11959/j.issn.2096-3750.2021.00221

Asbtract ( 152 )   HTML ( 26)   PDF (2323KB) ( 215 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

For the problem of strong randomness in space-time characteristics of wind speed in complex mountains，in order to improve the accuracy of wind speed data prediction, a neural network wind speed prediction algorithm based on multi prediction model and nonlinear combination was proposed.In the first layer of the algorithm, the grey wolf optimizer (GWO) and the dynamic convergence factor were used to improve the whale optimization algorithm (WOA), and the improved WOA was applied to the updating process of BPNN weights and bias items.At the same time, the improved whale optimiza-tion algorithm of back propagation neural network (IWOABP), ELM and LSTM three complementary single methods were constructed to build a combination prediction method, and on this basis, the ELM mixing mechanism of the second layer of the algorithm was utilized to learn the relationship between the first layer and the final result in a non-linear way.Simulation results show that compared with BPNN, WNN and GWOBP, the proposed algorithm has lower prediction errors.

Prediction of optimal growth parameters of barley seedling based on Kalman filter and multilayer perceptron

Yunlong HUANG, Zhengquan LI, Yujia SUN

2021, 5(4):  90-98.  doi:10.11959/j.issn.2096-3750.2021.00218

Asbtract ( 122 )   HTML ( 19)   PDF (1809KB) ( 231 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In order to improve the quality and planting efficiency of barley seedlings in the growth chamber, the Kalman filter algorithm was firstly used to process the data collected by the sensor, which effectively reduced the influence of environmental factors and the error of the sensor itself, improved the accuracy of the collected data, and ensured the precise control in the growth chamber and accurate test data.Then multiple nonlinear regression, radial basis function and multilayer perceptron neural network were used to analyze the average growth height, seedling weight and seed weight of barley seeds about 160 hours after germination under different conditions.The drying ratio was analyzed and compared.The results show that the multi-layer perceptron network model fits the data best.Using this model to predict the average height of barley seedlings and the ratio of seedling weight of barley seedlings in the optimal environment is basically consistent with the actual planting effect, which provides a certain reference for the planting of barley seedlings in the growth chamber.

Design and implementation of NB-IoT based environmental temperature and humidity monitoring system

Minmin MAO, Jiaqi JU, Yuling OUYANG, Yan JIN

2021, 5(4):  99-106.  doi:10.11959/j.issn.2096-3750.2021.00223

Asbtract ( 370 )   HTML ( 114)   PDF (2072KB) ( 286 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In view of the disadvantages of the traditional environmental monitoring system, such as high power consumption, high maintenance cost and no real-time monitoring, an environmental temperature and humidity monitoring system based on narrow band Internet of things (NB-IoT) technology was developed, which is mainly controlled by singlechip microcomputer.The work flow of the main controller and the working mode of the communication module were optimized to reduce the power consumption of the system.NB-IoT communication technology was utilized for data transmission to the cloud platform through the base station and core network.China Mobile IoT open platform (OneNET platform) was adopted to achieve the connection between the terminal device and the application, and expanded functions provided by the platform was used to design a mobile phone application to achieve the remote real-time monitoring of environmental parameters and improve the user’s convenience.

A survey on Wi-Fi signal based identification technology

Zhongcheng WEI, Xinqiu ZHANG, Bin LIAN, Wei WANG, Jijun ZHAO

2021, 5(4):  107-119.  doi:10.11959/j.issn.2096-3750.2021.00213

Asbtract ( 469 )   HTML ( 82)   PDF (1288KB) ( 756 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

With the constant progress of the information technology, the research on identification is developing towards unawareness and automation, and applied widely in many fields, such as human-computer interaction, smart home and intelligent security.As a non-contact wireless sensing technology, Wi-Fi signal based identification is progressively evolving towards the directions of multi-attribute identification and legality verification.Firstly, the principle of wireless signal-based identification was introduced, and reviews pretreatment processes from the viewpoint of signal processing were reviewed.Then, identification implementation was discussed from two different bases of walking and daily activities.Two major development trends of identification which were multi-attribute identification and legality verification were discussed from function evolution’s views.Finally, the multiple research areas were indicated including hyper-level feature extraction, daily activities’ analysis, identification model construction, and multi-technology fusion as future directions of realizing more robust, diversify and universal identification, which would provide the important support for the research and development of Internet of things perception layer.

Service and Application

Offshore aquaculture platform control system in intelligent fishery era

Jianbo ZHANG, Yu WANG, Xuejun NIE, Guoqing WU, Jiujun LIU, Jun YAN

2021, 5(4):  120-136.  doi:10.11959/j.issn.2096-3750.2021.00247

Asbtract ( 439 )   HTML ( 85)   PDF (7372KB) ( 294 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

The control system of offshore aquaculture platform was analyzed and the collaborative control technology with edge intelligence was introduced under the open architecture of cloud computing.A distributed intelligent control mode was established.A general framework architecture of cloud-edge collaborative control system for offshore aquaculture platform and an open architecture of intelligent aquaculture cloud platform were proposed exploratively, which could be deployed flexibly with different patterns of cloud-network-edge-terminal-intelligence.A development idea of control system and an implementation scheme of could central control and edge intelligence for offshore aquaculture platform were put forward.An attempt was made to build the theoretical framework and ideology system of cloud-edge collaborative control, and it was to construct the manner with the whole process of intelligent aquaculture and land-sea linkage operation and management mode for offshore fish-farm, with a view to finally realizing the intelligent fishery industrial ecology and green sustainable development goals.

Electricity meter area identification technology based on channel coding theory

Yinghai XIE, Yu ZHANG

2021, 5(4):  137-144.  doi:10.11959/j.issn.2096-3750.2021.00245

Asbtract ( 164 )   HTML ( 17)   PDF (1007KB) ( 375 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Station area identification has always been a problem in the ubiquitous power Internet of things, a new analytical model similar to fading channel was established for the station area identification, and based on the remote on-load voltage regulation function of on-load capacity and voltage regulation distribution transformer, the channel coding theory was applied to solve the problem.In order to identify station area of the smart meters, different special voltage regulation schemes with the properties of repeated coding and block coding were applied synchronously.At the same time, the power consumption information acquisition system was used to collect the voltage of each smart meter in this period.Finally, the recognition of each meter was realized based on the principle of minimum distance of decoding algorithm.The theoretical analysis and test results show that, compared with the traditional technology, the new technology can realize the high accuracy identification of a large number of meters in the on-load capacity and voltage regulation distribution transformer area in the existing power grid system without upgrading, which has high engineering application value.

Design and application of a lightweight management software for Internet of things node

Shiyou GUAN, Zaiqun WU

2021, 5(4):  145-152.  doi:10.11959/j.issn.2096-3750.2021.00248

Asbtract ( 207 )   HTML ( 46)   PDF (5423KB) ( 226 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

The modular design idea was used to abstract the logic of IoT nodes, and the function and control logic of nodes were divided into atoms, the atomic modules can be freely combined and deployed according to different application, they support single system and distributed system deployment, and were pluggable and easy to expand, meet the diversified design needs of IoT applications.The design and implementation of lightweight IoT node management software based on this idea were described in detail.The test results show that the software only needs 14 872 byte of flash and 1 976 byte of RAM, which can be used in limited hardware resources.This software framework was used to design smart hotel RCU(room control unit), including the design of voice gateway node based on esp32 (Wi-Fi SOC) and 4-bit key panel node based on cx32l003 (MCU).The RCU designed with this software framework can deploy to different hotels only needed to be designed once, which solves the problem that RCU cannot be used in different hotels and needs to be designed and modified repeatedly, and greatly speeds up the process of smart hotel deployment.At present, the RCU has been deployed and operated stably in Xi'an, Tupai and other express hotels.