System Architectures for Multi-Sensor Task Allocation

Heterogeneous sensor networks are increasingly being deployed to support users on the field who require many different kinds of sensing tasks, such as detecting people who may need help or monitoring collapsing buildings. Sensors may be scarce and in high demand. Sensing tasks might compete for the exclusive usage of available sensors. Such an environment is highly dynamic with, for example, sensor failures and changing sets of tasks. First responders usually lack the time and expertise to manually decide which are the best sensors for each task. We need therefore to automate the allocation of sensors to tasks. We formalize this problem, which we call Multi-Sensor Task Allocation (MSTA). We propose different system architectures to solve MSTA and we discuss the tradeoffs that each architectural choice implies. We show a proof-of-concept implementation of the system interface deployable on mobile devices in the field.