Fusing uncertain knowledge and evidence for maritime situational awareness via Markov Logic Networks

Abstract

The concepts of event and anomaly are important building blocks for developing a situational picture of the observed environment. We here relate these concepts to the JDL fusion model and demonstrate the power of Markov Logic Networks (MLNs) for encoding uncertain knowledge and compute inferences according to observed evidence. MLNs combine the expressive power of first-order logic and the probabilistic uncertainty management of Markov networks. \ud Within this framework, different types of knowledge (e.g. a priori, contextual) with associated uncertainty can be fused together for situation assessment by expressing unobservable complex events as a logical combination of simpler evidences. \ud We also develop a mechanism to evaluate the level of completion of complex events and show how, along with event probability, it could provide additional useful information to the operator.\ud Examples are demonstrated on two maritime scenarios of rules for event and anomaly detection

Context-based Information Fusion: A survey and discussion

“…The recent approach of Snidaro et al [98] discusses the fusion of uncertain sensory and contextual information for maritime situational awareness. Starting from the premise that events and anomalies are key elements in the process of assessing and understanding the observed environment, the paper arguments how building an effective situational picture for a surveillance system in the maritime domain involves combining high-level information with sensory data.…”

“…Even though almost all of the domains surveyed can be seen in terms of IF as soon as multiple sources of data/information are present and there is the need to combine their products in order to obtain better estimates of a certain variables, typical IF systems and applications generally have the common problem of lack of direct information from the focal entities of interest. SA systems, for example, have to go through a number of processing steps, also combining heterogeneous data, in order to estimate the status and intentions (or purpose) of non-cooperative entities (or process/system) [98]. In addition, observations from sensors are generally noisy and sources of information can have different level of trust and provide outputs with different quality [135], therefore making fusion a real necessity [124].…”

Context-based multi-level information fusion for harbor surveillance

“…Deductive reasoning is applied to detect inconsistencies between the situations obtained as a dynamic instantiation of the scene model and the situational patterns defined in the normalcy model. Normalcy rules are local to a navigational context, which depends in most cases on the geographical situation of the vessel (as in [14]). Inconsisten-cies denote abnormal situations that may indicate a potential threat.…”

“…A related proposal is presented in [14]. The authors use Mar-kov Logic Networks (MLNs) to represent uncertain context knowl-edge and automatically detect anomalies in the maritime domain.…”

“…1 To the best of our knowledge, this is the first attempt to combine ontologies and TBM-based uncertain reasoning to implement multi-level information fusion. Similar approaches in the literature have focused on alternative probabilistic models; namely, MultiEntity Bayesian Networks [12,13] and Markov Logic Networks [14]. Ontologies facilitate the creation of a computable model representing complex situational context (problem entities, scenario geometry, spatial relationships, etc.…”