Most recent issue published online in the International Journal of Critical Computer-Based Systems.

1. Detection of cyber-attacks for sensor measurement data using supervised machine learning models for modern power grid system
   The smart power grid systems are continually exposed to malicious cyber-attacks that are difficult to detect. If smart power grid attacks are not identified quickly and correctly, they may cause substantial economic losses and damage to the power system. To enhance productivity and improve the security of the smart power grid system against cyber-attacks, real-time detection of smart power grid attacks is still challenging. In recent years, there have been more cyberattacks, which have caused a lot of damage to power systems. This paper presents an experimental investigation of seven different approaches for detecting malicious activities and cyberattacks in the smart power grid system. Further, we employed maximum relevancy and minimum redundancy-hesitant fuzzy set feature selection technique to boost the attack detection performance. The experimental results demonstrate that random forest achieved the highest performance and average accuracy for two-class (95.30%) and three-class (95.33%) classifications, which shows that the presented proposed Model notably outperformed the other cyber-attack detection models.
2. An enhanced digital image watermarking technique using DWT-HD-SVD and deep convolutional neural network
   This paper proposes a novel image watermarking model, which combines discrete wavelet transform (DWT), Hessenberg decomposition (HD), singular value decomposition (SVD)-based deep convolutional neural networks (D-CNN) technique to explore the subjective and objective quality of the images. Initially, the source and cover image are preprocessed using random sampling techniques. During the process of embedding a watermark image, the cover image is decomposed into a number of sub-bands using the DWT process and the resulting coefficients are fed into the HD process. In continuation to it, the source image is operated on the SVD simultaneously and finally, the cover image is embedded into the source image by the attack-defending process. The probability of data loss during the watermarking extraction process and this issue is postulated by the D-CNN technique that explores the denoising process on the extracted watermarked images. The experimental results show that the proposed method has a good trade-off between robustness and invisibility even for the watermarks with multiple sizes.
3. Task models for mixed criticality systems - a review
   The past decade has seen tremendous interest in mixed criticality systems research due to its exponential growth with inherent challenges of effective resource utilisation and isolation. The pervasiveness of these systems along with their certification needs, prompt for suitable task models to perform the required analysis. Extensive usage scenarios and strict certification requirements have spawned a broad spectrum of research and evolved into several task models. In this work, a thematic survey of task models for both uni-core and multi-core mixed criticality systems is carried out. The work categorises task models based on attributes such as resources, quality of service, operating system overheads, energy, fault tolerance and parallel processing. After synthesising the state-of-the-art, the work summarises task models by providing a visual aid and a ready reckoner with traceability to mixed criticality challenges. This work serves as a quintessential reference manual for researchers and academicians in the mixed criticality domain.
4. Tuna swarm optimisation-based feature selection and deep multimodal-sequential-hierarchical progressive network for network intrusion detection approach
   Network intrusion detection system (NIDS) is important for securing network information. Neural network (NN) has recently been used for NIDS, which gained prominence results. Conventional neural network (CNN) has been introduced in network traffic data because of its single structure. The classification of assaults will no longer be useful due to redundant or inefficient features. Tuna swarm optimisation (TSO) has been introduced for feature selection (FS). First, pre-processing and feature extraction stages enable more efficient processing of features if handled independently. In order to examine the exploration space accuracy and position the best features, the second feature selection step of the TSO methodology involved selecting a subset of features by reducing the number of features. Lastly, multimodal deep auto encoder (MDAE) and gated recurrent unit (GRU) allow deep multimodal-sequential-hierarchical progressive network (DMS-HPN) intrusion detection method. Its DMS-HPN technique would routinely learn the temporal features among neighbouring network connections, simultaneously integrating diverse feature information inside a network. Datasets like UNSW-NB15 and CICIDS 2017 assess the effectiveness of the proposed DMS-HPN approach. Classification algorithms are evaluated via precision, recall, F-measure, and accuracy. Compared to conventional classifiers, the presented DMS-HPN classifier achieves the greatest accuracy.