Hamed Yaghoubi Shahir

Intelligent decision support for Marine safety and Security Operations

The architecture and core mechanisms of a decision support system for a Marine Security Operations Centre (MSOC) are presented. The goal of this system is to improve coordination in emergency response services during critical situations, including detection and prevention of illegal activities. The system design emphasizes robustness and scalability through its decentralized control structure, automated planning and replanning, dynamic resource configuration management and task execution management under uncertainty. An example scenario from the marine operations domain is described.

A Collaborative Decision Support Model for Marine Safety and Security Operations

Collaboration and self-organization are hallmarks of many biological systems. We present the design for an intelligent decision support system that employs these characteristics: it works through a collaborative, self-organizing network of intelligent agents. Developed for the realm of Marine Safety and Security, the goal of the system is to assist in the management of a complex array of resources in both a routine and emergency role. Notably, this system must be able to handle a dynamic environment and the existence of uncertainty. The decentralized control structure of a collaborative self-organizing system reinforces its adaptiveness, robustness and scalability in critical situations.

Anomaly detection in spatiotemporal data in the maritime domain

Maritime security is critical for many nations to address the vulnerability of their sea lanes, ports and harbours to a variety of threats and illegal activities. With increasing volume of spatiotemporal data, it is ever more problematic to analyze the enormous volume of data in real time. This paper explores a novel approach to representing spatiotemporal data for model-driven methods for detecting patterns of anomalous behaviour in spatiotemporal datasets.

Maritime situation analysis framework: Vessel interaction classification and anomaly detection

Maritime domain awareness is critical for protecting sea lanes, ports, harbors, offshore structures like oil and gas rigs and other types of critical infrastructure against common threats and illegal activities. Typical examples range from smuggling of drugs and weapons, human trafficking and piracy all the way to terror attacks. Limited surveillance resources constrain maritime domain awareness and compromise full security coverage at all times. This situation calls for innovative intelligent systems for interactive situation analysis to assist marine authorities and security personal in their routine surveillance operations. In this article, we propose a novel situation analysis approach to analyze marine traffic data and differentiate various scenarios of vessel engagement for the purpose of detecting anomalies of interest for marine vessels that operate over some period of time in relative proximity to each other. We consider such scenarios as probabilistic processes and analyze complex vessel trajectories using machine learning to model common patterns. Specifically, we represent patterns as left-to-right Hidden Markov Models and classify them using Support Vector Machines. To differentiate suspicious activities from unobjectionable behavior, we explore fusion of data and information, including kinematic features, geospatial features, contextual information and maritime domain knowledge. Our experimental evaluation shows the effectiveness of the proposed approach using comprehensive real-world vessel tracking data from coastal waters of North America.

Maritime Situation Analysis: A Multi-vessel Interaction and Anomaly Detection Framework

Maritime security is critical for protecting sea lanes, ports, harborsand other critical infrastructure against a broad range of threats and illegal activities like smuggling, human trafficking, piracy and terrorism. Limited resources constrain maritime domain awareness and compromise full security coverage at all times. This situation calls for innovative intelligent systems for interactive situation analysis to assist marine authorities and security personal in their routine surveillance operations. In this paper, we propose a novel situation analysis approach to analyze, detect and differentiate a range of interaction patterns and anomalies of interest for marine vessels that operate over some period of time in relative proximity to each other. We analyze vessel interaction scenarios to model common patterns as probabilistic processes in terms of hidden Markov models. To differentiate suspicious activities from unobjectionable behavior, we explore fusion of data and information from observable behavior (geospatial aspects, kinematic features and contextual information) and maritime domain knowledge from diverse sources. Our experimental evaluation using real-world vessel tracking data shows the effectiveness of the approach.

Maritime situation analysis

A methodical approach to analysis & design of computational situation analysis models based on model-driven engineering principles is proposed. Rendezvous anomaly detection in maritime safety and security serves as a realistic domain model to exemplify the problem scope and illustrate challenges and needs in developing practically working solutions. A novel approach to detecting anomalous multi-vessel interactions along with a generalized definition of rendezvous scenarios is presented.

Generating test cases for marine safety and security scenarios: a composition framework

In this paper we address the problem of testing complex computer models for infrastructure protection and emergency response based on detailed and realistic application scenarios using advanced computational methods and tools. Specifically, we focus here on testing situation analysis decision support models for marine safety & security operations as a sample application domain. Arguably, methodical approaches for analyzing and validating situation analysis methods, decision support models, and information fusion algorithms require realistic vignettes that describe in great detail how a situation unfolds over time depending on initial configurations, dynamic environmental conditions and uncertain operational aspects. Meaningful results from simulation runs require appropriate test cases, the production of which is in itself a complex activity. To simplify this task, we introduce here the conceptual design of a Vignette Generator that has been developed and tested in an industrial research project. We also propose a framework for composing vignettes from reusable vignette elements together with a formal representation for vignettes using the Abstract State Machine method and illustrate the approach by means of various practical examples.

Hidden Markov based anomaly detection for water supply systems

Considering the fact that fully immunizing critical infrastructure such as water supply or power grid systems against physical and cyberattacks is not feasible, it is crucial for every public or private sector to invigorate the detective, predictive, and preventive mechanisms to minimize the risk of disruptions, resource loss or damage. This paper proposes a methodical approach to situation analysis and anomaly detection in SCADA-based water supply systems. We model normal system behavior as a hierarchy of hidden semi-Markov models, forming the basis for detecting contextual anomalies of interest in SCADA data. Our experimental evaluation on real-world water supply system data emphasizes the efficacy of our method by significantly outperforming baseline methods.

Distributed Situation Analysis

Situation Analysis is critical for dynamic decision-making in responding to real-world situations. The complex and intricate nature of situation analysis processes calls for evolutionary modeling and formal engineering methods that facilitate experimental validation of abstract mathematical descriptions to link the essential design aspects with rapid prototyping in early development phases. For the transition from abstract concepts and requirements to precise specifications to high level design of situation analysis systems, we derive here a generic ASM ground model as a framework for defining the precise meaning of fundamental situation analysis concepts applicable to different application domain models.

Refactoring abstract state machine models

The Abstract State Machine (ASM) method proposes the concept of ground models for analyzing a target system based on pseudo-code-like descriptions for reasoning about system properties in terms of state machine runs over abstract data structures. This highly iterative process builds on stepwise refinement of ground models that evolve with progressing understanding of functional system requirements. Usually, as complexity increases, reorganization of a models internal structure helps enhance its flexibility and robustness. While this approach is common practice, the underlying principles are usually left implicit. In this paper, we propose refactoring patterns to restructure abstract machine models with the goal of improving their intelligibility and maintainability.