The increased use of Industrial Wireless Sensor Networks (IWSN) in a variety of different applications, including those that involve critical infrastructure, has meant that adequately protecting these systems has become a necessity. These cyber-physical systems improve the monitoring and control features of these systems but also introduce several security challenges. Intrusion detection is a convenient second line of defence in case of the failure of normal network security protocols. Anomaly detection is a branch of intrusion detection that is resource friendly and provides broader detection generality making it ideal for IWSN applications. These schemes can be used to detect abnormal changes in the environment where IWSNs are deployed. This paper presents a literature survey of the work done in the field in recent years focusing primarily on machine learning techniques. Major research gaps regarding the practical feasibility of these schemes are also identified from surveyed work and critical water infrastructure is discussed as a use case.
12 April 2019 in Human-centric Computing and Information Sciences
The threat to critical water system infrastructure has increased in recent years as is evident from the increasing number of reported attacks against these systems. Preventative security mechanisms are often not enough to keep attackers out so a second layer of security in the form of intrusion detection is paramount in order to limit the damage of successful attacks. In this paper several traditional anomaly detection techniques are evaluated in the context of attack detection in water distribution systems. These algorithms were centrally trained on the entire feature space and compared to multi-stage detection techniques that were designed to isolate both local and global anomalies. A novel ensemble technique that combines density-based and parametric algorithms was also developed and tested in the application environment. The traditional techniques had comparable results to the multi-stage systems and when used in conjunction with a local anomaly detector the performances of these algorithms were greatly improved. The developed ensemble technique also had promising results outperforming the density-based techniques and having comparable results to the parametric algorithms.
Search and rescue missions still typically rely on teams of people tirelessly searching for victims. Advances in technology have seen the use of drones to help the rescuers in their search. These drones primarily use visual systems but cost and visibility are major drawbacks with those systems. The system described in this paper attempts to resolve the localisation problem for drones used in search and rescue operations by making use of acoustic source localisation. The source location is estimated in terms of an azimuth estimation (the bearings or angle towards the source) as well as the distance. The proposed system showed promising results for both the azimuth and distance estimations.
Data integrity in wireless sensor networks (WSN) is very important because incorrect or missing values could result in the system making suboptimal or catastrophic decisions. Data imputation allows for a system to counteract the effect of data loss by substituting faulty or missing sensor values with system-defined virtual values. This paper proposes a virtual sensor system that uses multi-layer perceptrons (MLP) to impute sensor values in a WSN. The MLP was trained using a genetic algorithm which efficiently reached an optimal solution for each sensor node. The system was able to successfully identify and replace physical sensor nodes that were disconnected from the network with corresponding virtual sensors. The virtual sensors imputed values with very high accuracies when compared to the physical sensor values.
There has been an increasing number of attacks against critical water system infrastructure in recent years. This is largely due to the fact that these systems are heavily dependent on computer networks meaning that an attacker can use conventional techniques to penetrate this network which would give them access to the supervisory control and data acquisition (SCADA) system. The devastating impact of a successful attack in these critical infrastructure applications could be long-lasting with major social and financial implications. Intrusion detection systems are deployed as a secondary defence mechanism in case an attacker is able to bypass the systems preventative security mechanisms. In this thesis, behavioural intrusion detection is addressed in the context of detecting cyber-attacks in water distribution systems. A comparative analysis of various predictive neural network architectures is conducted and from this a novel voting-based ensemble technique is presented. Finally an analysis of how this approach to behavioural intrusion detection can be enhanced by both univariate and multivariate outlier detection techniques It was found that multiple algorithms working together are able to counteract their limitation to produce a more robust algorithm with improved results.
Wireless Sensor Networks (WSNs) are increasingly deployed in Internet of Things (IoT) systems for applications such as smart transportation, telemedicine, smart health monitoring and fall detection systems for the elderly people. Given that huge amount of data, vital and critical information can be exchanged between the different parts of a WSN, good management and protection schemes are needed to ensure an efficient and secure operation of the WSN. To ensure an efficient management of WSNs, the Software-Defined Wireless Sensor Network (SDWSN) paradigm has been recently introduced in the literature. In the same vein, Intrusion Detection Systems, have been used in the literature to safeguard the security of SDWSN-based IoTs. In this paper, three popular Artificial Intelligence techniques (Decision Tree, Naïve Bayes, and Deep Artificial Neural Network) are trained to be deployed as anomaly detectors in IDSs. It is shown that an IDS using the Decision Tree-based anomaly detector yields the best performances metrics both in the binary classification and in the multinomial classification. Additionally, it was found that an IDS using the Naïve Bayes-based anomaly detector was only adapted for binary classification of intrusions in low memory capacity SDWSN-based IoT (e.g., wearable fitness tracker). Moreover, new state-of-the-art accuracy (binary classification) and F-scores (multinomial classification) were achieved by introducing an end-to-end feature engineering scheme aimed at obtaining 118 features from the 41 features of the Network Security Laboratory-Knowledge Discovery in Databases (NSL-KDD) dataset. The state-of-the-art accuracy was pushed to 0.999777 using the Decision Tree-based anomaly detector. Finally, it was found that the Deep Artificial Neural Network should be expected to become the next default anomaly detector in the light of its current performance metrics and the increasing abundance of training data.