Modelling Cohesive-Frictional Particulate Solids for Bulk Handling Applications

Many powders and particulate solids are cohesive in nature and the strength often exhibits dependence on the consolidation stress. As a result, the stress history in the material leading up to a handling scenario needs to be considered when evaluating its handleability. This paper outlines the development of a DEM contact model accounting for plasticity and adhesion force, which is shown to be suitable for modelling the stress history dependent cohesive strength. The model was used to simulate the confined consolidation and the subsequent unconfined loading of iron ore fines with particle sizes up to 1.18mm. The predicted flow function was found to be comparable to the experimental results. 

Handling Sequential Observations in Intelligent Surveillance

Demand for intelligent surveillance in public transport systems is growing due to the increased threats of terrorist attack, vandalism and litigation. The aim of intelligent surveillance is in-time reaction to information received from various monitoring devices, especially CCTV systems. However, video analytic algorithms can only provide static assertions, whilst in reality, many related events happen in sequence and hence should be modeled sequentially. Moreover, analytic algorithms are error-prone, hence how to correct the sequential analytic results based on new evidence (external information or later sensing discovery) becomes an interesting issue. In this paper, we introduce a high-level sequential observation modeling framework which can support revision and update on new evidence. This framework adapts the situation calculus to deal with uncertainty from analytic results. The output of the framework can serve as a foundation for event composition. We demonstrate the significance and usefulness of our framework with a case study of a bus surveillance project.

Random Forest Based Approach for Concept-Drift Handling

Algorithms for concept drift handling are important for various applications including video analysis and smart grids. In this paper we present decision tree ensemble classication method based on the Random Forest algorithm for concept drift. The weighted majority voting ensemble aggregation rule is employed based on the ideas of Accuracy Weighted Ensemble (AWE) method. Base learner weight in our case is computed for each sample evaluation using base learners accuracy and intrinsic proximity measure of Random Forest. Our algorithm exploits both temporal weighting of samples and ensemble pruning as a forgetting strategy. We present results of empirical comparison of our method with îriginal random forest with incorporated replace-the-looser forgetting andother state-of-the-art concept-drift classiers like AWE2.