First line of defence

Outdoor perimeter intrusion detection sensors have been around for several decades now. Originally developed for use by military and Government agencies, outdoor sensors were – initially, at least – extremely expensive and out of reach of the commercial market. They were also quite unreliable because technology hadn't developed to the point where sensors could be counted on to work in varying weather and climatic conditions.

Outdoor sensors began to receive serious and widespread consideration in the general marketplace around a decade ago. Refinements (in particular concerning digital signal processing) continued to enhance performance both in detection and in the rejection of environmental and other non-threat sources. Costs also came down significantly, thereby allowing the commercialisation of sensor systems.

When it comes to procurement, security professionals can now choose from a range of intrusion detection sensor technologies that will address the entire spectrum of threat scenarios, environmental conditions, site constraints and nuisance alarm considerations.

Going back to basics
There are five basic elements to every outdoor perimeter protection system. First, it must deter. The perimeter should be designed in such a way as to clearly indicate the boundaries of the protected area while discouraging easy access to that area.

Second, the system must detect. When access to the protected area does occur, the outdoor perimeter sensor system provides early warning that unauthorised entry has indeed occurred.

Third, the system has to delay. Inherent in the detection process should be adequate delay to permit proper assessment of the intruder as well as delay in reaching critical assets. Then the detection system must assess and respond. The alarm has to be assessed to provide positive confirmation that it's a valid alarm. Based on that assessment, appropriate action can then be taken.

Before the security manager is able to select an appropriate sensor or sensor technology (often in conjunction with their consultant and/or installer), they should first understand the nature of the threat(s) and the potential sophistication of the intruder. Threat scenarios must be analysed and a clear distinction made between real and perceived threats. There must be a thorough understanding of site topography, and the availability of a 'clear zone' – or lack thereof – established.

In selecting a sensor technology or a specific product, the security professional should then look at performance with regard to several criteria (including the probability of detection, the vulnerability to defeat, the false alarm rate and the nuisance alarm rate). These criteria equate to strengths and weaknesses, and understanding them is vital to selecting the right technology for the application.

The probability of detection is the likelihood (expressed as a percentage) that the chosen sensor will recognise an intruder within the detection zone. A given sensor's probability of detection is conditional upon target characteristics (such as mass, size, speed and orientation), sensitivity settings, weather conditions, site conditions and proper maintenance of the equipment. A probability of detection of 100% may be achieved by making the sensor so responsive that the nuisance alarm rate becomes exceptionally high. End users should be aware that the probability of detection for outdoor sensors is typically specified at 99%, with anything less than 95% considered unacceptable. A probability of detection of 95% means that one-in-20 intrusions will not be detected at all.

Vulnerability to defeat is the likelihood that a sensor could be beaten (for example due to bypass or spoofing). Each sensor technology will have different vulnerabilities inherent in its design. The vulnerability to defeat will also depend on the intruder's foreknowledge of the perimeter protection system, the availability of aids such as ladders and the operational readiness and state of repair of the system.

False and nuisance alarms
Vulnerability to defeat may be minimised by end users if they deploy covert (ie hidden) sensors by combining sensor technologies for high risk sites and eliminating weaker points in the perimeter protection coverage.

The false alarm rate is a term used to describe the rate of invalid alarms caused by unknown sources. Typically, the cause is the sensor's own self-generated noise, but it can also include temperature extremes above or below what the sensor's electronics were designed to withstand, poor grounding, faulty or improper electrical connections, power supply voltage fluctuations, damage to equipment perpetrated during installation and poor sensor maintenance. The false alarm rate is expressed as a number per day, month or year per sensor. It's normally negligible, and can range from one per month for low end sensors to one per year for high end set-ups.

In a similar vein, the nuisance alarm rate is the rate of invalid alarms caused by identifiable non-threat sources, normally expressed as a number per day or month per sensor. It differs from the false alarm rate in that the source of the alarm can be readily identified by the system operator (typically by way of remotely-monitored CCTV).

This particular alarm rate is most commonly associated with small animals, environmental influences (such as heavy winds, rain, blowing snow, fog, hail and day/night, etc), terrain conditions, vegetation and non-threat personnel in the detection zone.

The nuisance alarm rate will vary considerably depending on technology, the application and prevailing site conditions. A typical high end sensor will have a nuisance alarm rate of one per month, whereas low end sensors can have several per day.

It's really the responsibility of the security manager to determine what's an acceptable nuisance alarm rate for a particular application. A rate of zero will not be achievable in any application, while a nuisance alarm rate that's too high leads to a lack of confidence in the sensor (resulting in the sensor being turned off or the alarms ignored).

Proper management of a given system's false and nuisance alarm rates is critical to the overall effectiveness of your perimeter protection system. Without any thorough assessment, these invalid alarms can result in an unacceptably high dispatch rate and a security system that's either totally ignored or shut down by its human monitors.

Fortunately for end users, various techniques for managing invalid alarm rates have now been developed. These include selecting the proper technology for the application, supplementing intrusion detection sensors with CCTV for the remote assessment of each event, paying proper attention to site specifics, following the recommendations of the manufacturer and providing sufficient operation and maintenance support and budget for both the system and the site.

Comparison by category
Sensor technologies can be compared in at least five categories, namely operation, installation, configuration, detection zone type and application.

Operation: Intrusion detection sensors are either active or passive in nature. For their part, active sensors transmit energy and will detect any change in received energy (eg radio frequency, infrared or microwave). Passive sensors are designed to detect energy generated by an intruder or any changes in natural energy caused by the same person(s) (for instance vibration, heat or pressure).

Installation: Perimeter protection sensors may be visible or covert (hidden). Visible sensors can act as a deterrent and are typically easier to install and maintain, but are far more susceptible to vandalism and defeat. Covert sensors, meanwhile, are more difficult to detect and defeat and far less susceptible to acts of vandalism.

Configuration: Sensors are line-of-sight or terrain-following. Line-of-sight sensors require that the surrounding terrain be a clear, flat surface. Often, site preparation and maintenance will be required. Note that terrain-following sensors are able to detect equally well on either even or uneven terrain.

Detection zone type: The detection zone is either line or volumetric. The line detection zone is confined to a well-defined line or plane, is often visible and (generally speaking) easier to locate. It requires much less clear area than volumetric sensors which are designed to detect intruders in a given volume of space that can be both above and/or below ground. In general, the detection volume isn't visible and is thus more difficult for the intruder to locate.

Application: Sensors may be free-standing, fence-mounted or buried. Free-standing sensors are usually self-supported or mounted to free-standing poles or similar structures that effectively create the detection area. Fence-mounted sensors are mounted either directly to the fence or on stand-offs attached to it.

Ported coax sensors create an electromagnetic detection field around buried 'leaky' coaxial cables. Here, the field extends both above and below ground. The sensor – which operates in the VHF band – looks specifically for the mass, speed and conductivity characteristics of human intruders and vehicles alike while discounting the movements of small animals.

This type of sensor provides a wide detection field for the end user, and may be installed in sand, clay, gravel, frozen ground, asphalt or concrete. The technology is best classified as active, covert, terrain-following, volumetric and buried.

The only consistent contributors to nuisance alarm rates are moving surface water and metal objects in the detection zone. Interestingly, vulnerability to defeat for ported coax sensors is considered to be the lowest of all the sensor technologies currently available.

For their part, fence disturbance sensors come in a variety of forms, including motion sensing (vibration), acoustic and fibre optic. All types are mounted on the fence fabric and trigger an alarm when attempts are made to cut, climb or lift that fabric. These types of sensor are classified as passive, visible, terrain-following, line and fence-mounted. Acoustic is currently the most popular variant, motion sensing the least popular.

Recent developments include the introduction of digital signal processing technology to enhance detection and minimise nuisance alarm rates, improvements in the sensing cable to include power and data distribution and the ability to display the intrusion location within three metres.

Wind, hail, rain, snow, animals, scattering rubbish and vegetation hitting the perimeter fencing can all affect the nuisance alarm rate. This technology is considered to have a medium level vulnerability to defeat.