International Journal of Sensor Networks (5 papers in press)

Regular Issues

• Residual Spatial-Temporal Graph Convolutional Neural Network for On-Street Parking Availability Prediction  
  by Guanlin Chen, Sheng Zhang, Wenyong Weng, Wujian Yang 
  Abstract: Smart cities can provide people with a wealth of information to make their lives more convenient. Among many other benefits, effective parking availability prediction is essential as it can improve the overall efficiency of parking and significantly reduce city congestion and pollution. In this paper, we propose a novel model for parking availability prediction, i.e., the residual spatial-temporal graph convolutional neural network, which enhances the accuracy and efficiency of the prediction process. The model utilises graph neural networks and temporal convolutional networks to capture the spatial and temporal features, respectively, fusing through a residual structure called the residual spatial-temporal convolutional block. We conducted experiments using real-world datasets to compare the performance of the proposed model with that of the baseline models. The experimental results demonstrate that our model outperforms the baseline models in predicting the long-term parking occupancy rate and achieves the fastest prediction speed.
  Keywords: RST-GCNN; on-street parking availability prediction; graph neural network.
  DOI: 10.1504/IJSNET.2023.10058808
   
  
• NASA Space Station Rolling Bearings Anomaly Detection Based on PARA-LSTM Model  
  by Yingqian Zhang, Jiaye Wu, Hui Xie, Rongru Hua, Qiang Li 
  Abstract: Anomaly detection in time series data identifies abnormal events or behaviours. Traditional methods include principal component analysis (PCA) combined with Mahalanobis distance and long short-term memory (LSTM). Autoencoders and neural network techniques have been applied to the problem of anomaly detection. Still, challenges remain, such as large training data volume, network parameter initialisation, low training efficiency, and poor anomaly detection performance. This paper proposes an anomaly detection method based on parallel-long short-term memory (PARA-LSTM), which constructs two parallel processing structures. The method was tested on the rolling bearing vibration dataset collected by the NASA space station. It could detect anomalies five days ahead of the actual system destruction time, outperforming the PCA method by detecting anomalies one day earlier. PARA-LSTM has good performance, stability, and generalisation ability.
  Keywords: autoencoder; bearing vibration; anomaly detection; Mahalanobis distance; autoencoder network; parallel-long short-term memory; PARA-LSTM.
  DOI: 10.1504/IJSNET.2023.10060575
   
  
• Progressive Moisture Prediction Technique using Regressive Learning using Soil and Vegetation Data  
  by Xuetao Jia, Ying Huang 
  Abstract: Predicting soil moisture is crucial for optimal crop planting and improved yields across varying climates and soil types. Recent intelligent computing trends have enabled machine learning methods to predict soil moisture. This article introduces a progressive moisture prediction technique (PMPT) using regressive learning (RL) to predict soil moisture for precision farming. PMPT addresses prediction error from missing soil and vegetation sensor data. RL identifies and estimates missing data based on previous predictions. Prediction error margins are identified between vegetation yields using linear progression. PMPT is validated based on prediction values close to the error margin and accurate crop yield outputs. Factors like climate and maximum moisture periods are incorporated to compute minimum and maximum values around crop cycles. Thus, RL differentiates accurate and erroneous moisture detection from linear soil inputs regardless of data availability. PMPT is validated on metrics including prediction accuracy, error, time, differentiation, and saturation point.
  Keywords: climatic changes; moisture prediction; regressive learning; soil data; vegetation.
  DOI: 10.1504/IJSNET.2023.10060635
   
  
• Performance Analysis of Various Machine Learning Models for Membership Inference Attack  
  by K. Karthikeyan, K. Padmanaban, D. Kavitha, Jampani Chandra Sekhar 
  Abstract: In order to function correctly during the training phase, many ML models require enormous amounts of labelled data. There is a possibility that the data will contain private information, which must be protected regarding privacy. Membership inference attacks (MIA) are attacks that try to identify if a target data point was utilised for training a particular ML method. These attacks have the potential to compromise users’ privacy and security. The degree to which an algorithm for ML divulges user membership information varies from implementation to implementation. Hence, a performance analysis was performed based on different ML algorithms under MIA inference attacks. This study proposed for comparing different ML approaches against MIAs and analyses which ML algorithm is better performing to such privacy attacks. Based on the performance analysis observation, the GAN and DNN models are considered as the best ML models to defend against MIA attacks with better performances.
  Keywords: data acquisition; data security; inference attacks; MIA; machine learning; ML; pre-processing and privacy.
  DOI: 10.1504/IJSNET.2023.10060646
   
  
• LSTM based Multi-PIR Sensor Information Fusing for Estimating the Speed and Position of Pedestrians on Green Campus  
  by Yong Zhang, Lixin Zhao, Yansong Fang, Daoming Mu, Haiyan Zhang 
  Abstract: Pyroelectric infrared (PIR) sensors are popular for pedestrian detection in green campus lighting systems due to their low power consumption, low cost, and ease of installation. However, these sensors still fail to detect important information such as pedestrians’ speed and position, which are necessary for precise lighting control. This paper proposes a method to detect pedestrian speed and position by utilising fused information from network-connected PIR sensors. First, we collect and analyse the time-domain signal of the PIR sensor, aggregate the collected data according to the peak time sequence features, and perform feature enhancement operations on the collected data. The speed and position of pedestrians are then estimated by using the long short-term memory neural network. Finally, the effectiveness of the proposed method is validated through simulations and experiments, as demonstrated through improved success rates in identifying pedestrians’ moving position and speed, satisfying real-time detection requirements for green campus management.
  Keywords: pyroelectric infrared; PIR; sensors; pedestrian detection; green campus lighting systems; pedestrian speed; pedestrian position; fused information.
  DOI: 10.1504/IJSNET.2023.10060690