Performance measures for airborne maritime surveillance radars

A typical performance specification for an airborne maritime surveillance radar might be defined as the ability to detect, say, a 5 m/sup 2/ target in sea state 3 at a range of at least 25 nmi in clear air, with a detection probability of 0.9 and with one false alarm per minute. This is a basic statement of a user's needs but it is very inadequate as a specification of the radar performance that is actually required. The performance of an operational radar is characterised by many other features, which are not simply summarised in terms of a detection range and false alarm rate in a generalised environment. In particular, concepts such as probability of detection and probability of false alarm are conventionally steady state performance measures, for operation in thermal noise or uniformly distributed clutter. In practice, the radar will be dynamically adjusting its thresholds in response to a spatially and temporally varying environment. This will lead to a dynamic variation in detection and false alarm rate that cannot be simply described by traditional methods. This paper describes some of the problems with standard performance measures which are used for radar and proposes a number of further measures which should be considered to fully characterise a radar, especially under dynamically changing conditions.