Automated Real-Time Event Detection and Classification

IFOS proposes a software solution to real-world applications that require automated real-time event detection and classification based on multiple sensor sources. This software will integrate various data sources and extract features to provide a more accurate description of the event. In Phase I, the goal is to implement such software and prove its feasibility of detecting and classifying real-world events. The algorithms must be able to identify these events within a noisy and complicated background. Since the detection needs of the different potential applications are so diverse, different algorithms may need to be applied to the same data streams in order to detect different classes of features. The features extracted from individual data source and algorithm will require the integration. All of these features will need to be integrated together to achieve a robust, stable detection of real-world events. One potential application is to improve the spectral analysis with data sources from multiple spectral bandpass filters that capture a mixture of spectral signatures from the target. IFOS will identify more applications required by MDA and provide optimized system design based on findings. These applications may include multi spectral analysis, robotic control, Unmanned Aerial Vehicle (UAV) and Unmanned Ground Vehicle (UGCV).