Image sensors deployed in the Internet of Things (IoT) generate vast volumes of digital images. These images may be subject to deliberate alteration, compromising their trustworthiness. Estimating the trustworthiness of this image data is crucial for many applications; however, this aspect has not been adequately explored in the existing literature. In this article, we propose a robust and real-time trust estimation framework for IoT image data, leveraging numeric data generated from other types of sensors deployed in the same Area of Interest (AoI). The theoretical model was developed using statistical approaches, and Shannon’s entropy was employed to measure the uncertainty associated with sensor readings during a specific event. Later, we applied Dempster-Shafer theory (DST) of combination to fuse information collected from image as well as numeric data-generating sensors where both types of sensors were observing the same event in the same AoI concomitantly. To evaluate the proposed framework, we implemented an IoT testbed using LoRa sensor nodes, edge devices, an LoRaWAN gateway, the things network (TTN), and a data analytics server. The testbed was used to collect observation data of a fire event using image and temperature sensors in an indoor residential setup in different conditions. Consequently, eight data sets (four authentic and four hacked) were built, each containing both image and temperature data readings under various scenarios. The proposed trust framework accurately estimated the trust score of images (91% overall accuracy) across all the data sets and outperformed existing trust models.

We have developed a novel ’Teacher-Student with human feedback’ model for Human-Artificial Intelligence (AI) collaborations in cybersecurity tasks. In our model, AI furnishes sufficient information about its decision-making process to enable human agents to provide feedback to improve the model. Our key innovations include: enhancing the interpretability of AI models by analyzing falsely detected samples using LIME and SHAP values; developing a novel posthoc explanation-based dynamic teacher-student model to address concept drift or concept shift; integrating human experts’ feedback on falsely detected samples to increase accuracy, precision, and recall values, without retraining the entire model; establishing a list of attack-based feature values for human experts to promote reproducibility. We show in experiments with real data and threat detection tasks that our model significantly improves the accuracy of existing AI algorithms for these tasks.

Securing cloud environments such as Microsoft Azure cloud is challenging and vulnerabilities due to misconfigurations, especially with user roles assignment, are common. There have been significant efforts to find vulnerabilities that enable lateral movements in Azure AD systems. All of the existing works, however, either follow a manual process to find new vulnerabilities or are only able to discover whether known vulnerabilities exist in a deployed Azure environment. We develop an Azure Active Directory (AAD) lateral movement-discovery tool, CoZure, that can help researchers find new lateral movements in an Azure AD environment. CoZure deploys algorithms from Context-Free Grammar (CFG) to first learn the ways (grammar rules) that security researchers find vulnerabilities and then extend these rules to discover new lateral movement paths. CoZure first collects a large set of existing AAD environment commands using a specialized scraping tool, it then uses CFG to build a knowledge base dataset from these commands and previous attacks. Cozure then applies the knowledge learned to find new combinations of commands that could open up new candidate lateral movements, which are then tested in a real AD environment for validation and manually checked by the user. CoZure helped discover lateral movements that current fuzzing tools (e.g., OneFuzz, RESTler) cannot identify and also shows better performance in finding existing misconfiguration issues in Azure AD. Using CoZure, we have discovered two new (not previously known) lateral movement methods that could lead to numerous new attacking paths in Azure AD.

Smart home applications are becoming increasingly popular due to their ability to provide safety, comfort, and remote assistance. These applications are usually controlled using a smart home controller, which is often the target of malware attacks. A successful attack may result in financial loss, disclosure of personal and/or sensitive information, or even loss of human lives. Although existing research has employed machine learning models to detect various malware attacks in smart home systems, they haven’t directly tackled the issue of class imbalance in this domain. In addition, the use of ensemble learners is expected to provide improved performance. To address this, we investigated different oversampling techniques to increase the number of samples in the minority classes and incorporated ensemble learners to see their impact on the prediction performance. Experimental evaluation indicates a marked enhancement of 4-5% across metrics, encompassing accuracy, precision, recall, and the F-1 score.

To enhance industrial production and automation, rapid and faster transportation of raw materials and finished products to and from distributed factories, warehouses and outlets are essential. To reduce cost with increased efficiency, this will increasingly see the use of connected and self-driving commercial vehicles fitted with industrial grade sensors on roads, shared with normal and self-driving passenger vehicles. For its wide adoption, the trustworthiness of self-driving vehicles in the intelligent transportation system (ITS) is pivotal. In this article, we introduce a novel model to measure the overall trustworthiness of a self-driving vehicle considering on-Board unit (OBU) components, GPS data and safety messages. In calculating the trustworthiness of individual OBU components, CertainLogic and beta distribution function (BDF) are used. Those trust values are fused using both the dempster-Shafer Theory (DST) and a logical operator of CertainLogic. Results of our simulation show that our proposed method can effectively determine the trust of self-driving vehicles.

The increasing adoption of driverless cars already providing a shift to move away from traditional transportation systems to automated ones in many industrial and commercial applications. Recent research has justified that driverless vehicles will considerably reduce traffic congestions, accidents, carbon emissions, and enhance the accessibility of driving to wider cross-section of people and lifestyle choices. However, at present, people's main concerns are about its privacy and security. Since traditional protocol layers based security mechanisms are not so effective for a distributed system, trust value-based security mechanisms, a type of pervasive security, are appearing as popular and promising techniques. A few statistical non-learning based models for measuring the trust level of a driverless are available in the current literature. These are not so effective because of not being able to capture the extremely distributed, dynamic, and complex nature of the traffic systems. To bridge this research gap, in this paper, for the first time, we propose two deep learning-based models that measure the trustworthiness of a driverless car and its major On-Board Unit (OBU) components. The second model also determines its OBU components that were breached during the driving operation. Results produced using real and simulated traffic data demonstrate that our proposed DNN based deep learning models outperform other machine learning models in assessing the trustworthiness of individual car as well as its OBU components. The average precision of detection accuracies for the car, LiDAR, camera, and radar are 0.99, 0.96, 0.81, and 0.83, respectively, which indicates the potential real-life application of our models in assessing the trust level of a driverless car.