Intelligent task scheduling in sensor networks

Ever more complex sensors have become available to create and maintain situational awareness during missions. Choosing the most suited sensor for the execution of a sensor function is based on sensor capabilities and function attributes. When these characteristics change rapidly, sensor allocation for sensor functions will shift. To increase performance of the entire sensor network, the total set of sensors should be scheduled in a single system. This paper puts forward and compares three new methods for scheduling prioritised tasks in sensor networks. The first is based on fuzzy Lyapunov synthesis. The other two use a genetic algorithm (GA) to optimize the set of schedules. The second scheduler uses this set to (re)train a neural network to be used online. The third approach is a novel online use of the GA. Tests showed that the novel online GA leads to a robust scheduling algorithm with high overall performance.