Multi-Sensor Data Fusion and Smart Decision Making

Future intelligent systems will involve the deployment of large-scale sensor networks in which hundreds or thousands of microsensors, unexpensive, small and lightweight devices with integrated sensing, computing, communication and possibly actuation capabilities will work together to achieve a common mission-specific task. The nodes may be heterogeneous, with varying resources and mobility (some can be fixed, others mobile). The quality of service (QoS) requirements are diverse ranging from time-critical data to support robotics, fire control to background data, saving people in earthquake conditions to military operations in unknown and changeable environment. Such sensor networks have extended their territory to industrial applications, such as energy management, agricultural monitoring, for measuring temperature, sound, pressure, motion, pollutants, for process control and surveillance. Sensor networks will monitor the environment at close range with high special and temporal resolutions and will likely reveal previously unobservable phenomena in the physical world. This project will investigate novel probabilistic methods for extracting knowledge from data and decision making.