Artech author Kenneth Hintz gave us insight into theories of sensor management:
There is no single theory of sensor management due to the heterogeneity of sensors (physical, social, cyber, database), applicable mathematics (game theory, estimation, information, statistics, probability), and users with different needs (strategic, tactical) and values (local, national, political, ethical). What started out as the human observation of raw data from individual sensors dedicated to a single task on a single platform managed by a single user, has evolved over the last 50 years to networks of dissimilar sensors with diverse tasking on separate platforms which are used to satisfy the needs of a multiplicity of users. The transition has been from a human-in-the-loop (HIL) scheduling approach to a human-on-the-loop (HOL) management approach to controlling sensors. HOL requires a closed-loop autonomous sensor management system which is indirectly controlled and which has a measure of performance which transcends individual users and individual sensors and seeks to maximize a mission objective function. Furthermore, this management approach does not have the luxury of non-real-time optimization but rather requires real-time computation of the best next collection opportunity (BNCO).
To develop such a system, I looked to communications theory and utilized my experience first as an electronic warfare aircraft commander in the US Navy, and subsequent experience designing and deploying special collection systems (AN/ULQ-16) to develop a comprehensive sensor management objective function, the maximization of the expected information value rate (EIVR) which can be applied to the separable problems of choosing what data to collect and choosing which sensor to utilize in order to acquire that data.
Shannon’s well-known communications theory is predicated on maximizing information transfer through a bandwidth limited channel with known signal to noise characteristics. He was notably agnostic about what data to transmit through that channel. The complement to this approach which is applicable to sensor management is to assume that the channel (the sensor) is already designed to maximize information flow through it and the criterion to optimize is the transfer of information from the real environment to the mathematical model of that environment, i.e., situation assessment. The optimization criteria that was developed and presented in Sensor Management in ISR is the expected information value rate (EIVR). All four of the components of EIVR are real-time computable and combined in a six-component decomposition of the information based sensor management (IBSM) approach to sensor management. The result is that EIVR is a dimensionally conformal measure designed to minimize the mission valued uncertainty and maximize the actionable intelligence in the mathematical model of the situation. It is this probabilistic model, or knowledge, which is presented to the decision makers.
For more information or to order, click here.