An example of an advanced security system capable of detecting, tracking, and zooming in on intrusions (Photo Source: Nuclear Street) 

There are three types of performance measures used to describe sensing equipment. These are the probability of sensing (PS), the  Nuisance Alarm Rate (NAR), and the vulnerability to defeat.

Probability of Sensing

The probability of sensing (PS) is the probability that a sensor will alarm for a particular type of intrusion under certain conditions.  The PS value is a measured value and should be associated with a statistical confidence level. The design of detection systems should take into account certain factors that affect PS, such as environmental conditions (e.g., fog, rain, snow, etc.), installation conditions, and the condition of the equipment (a sensor's PS value may degrade over time). In addition, the design needs to account for the type of action by an intruder to be detected (e.g., running, walking, jumping, crawling, tunneling, breaching, etc.).  

For these reasons, the specific requirements of the sensor need to be established upfront. In addition, because detectors will have varying detection probabilities for different environments and different actions, more than one detector may be required to provide adequate coverage of an area. Typically, the best approach is to use multiple complementary sensors.  This will ensure adequate detection to the full range of actions that a potential intruder may take.

When designing the PPS, there may not be a sensor that provides an adequate PS. When this happens, redundant sensors can be installed.

When multiple complementary sensors are installed in one location, we need to calculate the total probability of sensing for the ensemble of sensors.  We do this by finding the total probability of non-detection for the group, multiplying together the complements for all of the sensor probabilities, and then subtracting this number from one to get the probability of sensing for the grouping of sensors.  So, for an ensemble of 3 complementary sensors, the total Probability of Sensing would be calculated as follows:

Total P_S = 1 - [(1 - P_S1) * (1 - P_S2) * (1 - P_S3)]

Where P_S1 is the P_S of sensor 1, P_S2 is the P_S of sensor 2, etc.

Nuisance Alarm Rate

The Nuisance Alarm Rate (NAR) is the number of alarms that a detector signals over some period of time that are not caused by an intrusion. These can be caused by sources external to the detector (such as animals, rain, wind, etc.) or by false alarms generated by the equipment itself (such as component failures or inadequate maintenance). If a sensor has a high NAR, it will repeatedly generate false alarms, which could lead to it eventually being ignored and to the potential diverting of resources away from other areas where they might actually be needed.

Vulnerability to Defeat

Any sensor can be defeated,and every sensor will have specific vulnerabilities that vary in complexity. There are two general ways to defeat a sensor. The first is by bypassing the sensor, which simply involves going outside of the sensor's detection volume. The second way is spoofing, which allows the adversary to traverse the sensor's detection volume without generating an alarm.