What security professionals need to understand about layered detection, nuisance alarms, and sensor fusion
When a perimeter alarm reaches your security team, you need to know what's happening, where it's happening, and whether there's time to respond. Repeated nuisance alarms make that job harder. So does a system that filters out activity your team needed to see.
That challenge was at the center of Jeff Drews’s presentation at Security Technology Forum 2026. Jeff, Director of Technical Solutions at Southwest Microwave, explored how perimeter detection is evolving in his session, “Beyond the Fence: Technology and the Evolution of the Perimeter Security Market.”
His message gives security leaders, consultants, and integrators a useful way to evaluate their next project: better processing and analytics create new possibilities, but effective protection still depends on the threat, the site, and the decisions that turn sensor data into an actionable alarm.
Jeff grounded the discussion in five familiar principles: deter, detect, delay, assess, and respond. A fence establishes a boundary and can help delay an intruder. Detection provides an alert. Assessment helps determine what is happening, and a response plan connects that information to action.
The design starts with understanding who might challenge the perimeter, what they are trying to reach, and what the consequences could be. An opportunistic trespasser and a prepared intruder targeting a critical asset present different problems. Those differences should influence sensor selection, detection coverage, and the number of protective layers.
Layers can extend beyond the property boundary, continue along a fence or wall, and protect individual assets inside the facility. Each layer needs a defined purpose and an appropriate detection technology.
Cameras are central to that plan. Drews emphasized the importance of pairing perimeter sensors with visual assessment so operators can investigate an alarm and understand the developing situation. Detecting movement is only part of the job; the team also needs usable information to guide its response.
A published false alarm rate can look reassuring while saying less than a buyer expects about everyday operations.
Jeff explained a distinction used in manufacturer specifications: a false alarm can mean an alarm generated without an external stimulus, while a nuisance alarm results from a stimulus that does not represent the intended intrusion threat. Weather, wildlife, vegetation, and other site activity can contribute to nuisance alarms.
The terminology matters when comparing products. A very low specified false alarm rate does not, by itself, establish how frequently a security team will investigate unwanted alarms after installation.
Ask what the manufacturer counts, how the system was tested, and what conditions apply. Then connect those answers to the workload at your site. Repeated nuisance alarms can erode operator confidence and contribute to alarm fatigue. Reducing that burden matters, provided the system continues to detect the threats it was installed to find.
Much of the evolution Jeff described comes from improved signal processing, diagnostics, connectivity, and the ability to interpret activity. Established sensing methods remain useful, while newer implementations provide more information and greater control over detection behavior.
Microwave detection remains part of the layered protection strategies Jeff discussed. Advances in processing and diagnostics can improve how those systems are configured and supported. Video has also gained more capable analytics that help interpret activity and support assessment.
The practical question is which capability improves the project. Location information, event history, and clearer diagnostics may each solve a different operational problem.
Distributed acoustic sensing, or DAS, uses fiber-optic cable as a sensing element. Analysis of disturbances along the cable can help detect and locate activity, with applications that include fence-mounted and buried perimeter detection.
Jeff showed how cleaner signals and improved analysis can help separate intrusion-related activity from background disturbance. In one example, he discussed distinguishing digging activity from nearby animal movement. The value is in giving the system better information to evaluate, with configuration suited to the installation.
Jeff also compared a two-dimensional LiDAR sensing plane with three-dimensional monitoring of a volume. The 3D system he demonstrated allowed detection zones and exclusion zones to be configured around the space being protected.
He described a short field test at a nuclear facility where vibration, fences, and water complicated detection. Carefully configured zones enabled the tested crawling, walking, and running movements to be detected without nuisance alarms during that two-day evaluation.
That example illustrates the value of matching configuration to a difficult application. Its results describe that particular test; performance at another site still needs to be established.
Classifying a target as a person, animal, or vehicle can help a security team focus on relevant activity. It also adds a timing requirement to the design.
Jeff explained that a system needs enough information to make its classification. If an alarm depends on identifying a person, someone moving quickly through a narrow zone may leave before the system completes that decision.
Zone dimensions, target speed, target presentation, and the selected alarm rules all matter. A configuration that performs well with a person walking upright needs additional testing for the other movements the site must detect.
This is also where the environment becomes part of the design. Optical sensors can be affected by fog, precipitation, and airborne particles. Jeff highlighted standing water and movement within the detection area as considerations for microwave installations. Each technology needs to be evaluated against its own limitations.
Sensor fusion can combine information from cameras, LiDAR, microwave detectors, and other devices to improve situational awareness. It can help evaluate early warning activity and reduce the nuisance alarms that consume an operator’s attention.
Jeff emphasized that the value and risk depend on how the information is combined.
“The fusion decision rule matters just as much as the sensors themselves.”
Consider a microwave sensor and a LiDAR sensor monitoring the same area. The alarm behavior changes depending on the logic:
| Alarm logic | What triggers an alarm | What to evaluate |
|---|---|---|
| AND | Both sensors must detect the event. | A valid detection by one sensor may be withheld if the other does not detect it. |
| OR | Either sensor can trigger an alarm. | Independent detections are retained, and the team must assess alarms from either source. |
For example, if environmental conditions reduce LiDAR performance while the microwave sensor detects an intrusion, an AND rule could prevent that detection from reaching the operator as an alarm.
This does not describe every form of sensor fusion. It demonstrates why the decision rule needs to be understood and tested. Combining location, classification, and video information can support assessment while preserving independent alarms.
For the high security applications he emphasized, Jeff favored that approach. The design should also address simultaneous targets, sensor degradation, and the information an operator receives when one device cannot confirm another device’s detection.
A successful demonstration is a starting point. The next step is to establish how the proposed configuration performs at the actual site.
Drews urged security professionals to test intrusion detection during relevant weather and operating conditions, alongside nuisance alarm performance. A system producing few alarms has not yet demonstrated that it reliably detects the required threats.
Use these six questions to guide project discussions:
For existing installations, Drews also noted that some upgrades may reuse wiring, power, and mounting infrastructure while replacing sensor heads. Compatibility is model-specific, and modern capabilities cannot automatically be added to an older system through firmware.
The session gives security professionals a practical way to evaluate the growing range of perimeter technologies. Start with the threat, give each detection layer a clear role, and understand how the system decides what reaches the operator.
Watch the full presentation for Drews’s demonstrations and discussion of fiber sensing, 3D LiDAR, and sensor fusion. Then bring those questions into your next design review.
Planning a perimeter upgrade or working through a challenging detection application? Connect with the MidChes team to discuss your site, your existing systems, and the options for strengthening detection and assessment.
A perimeter intrusion detection system, or PIDS, uses sensors to detect activity that may indicate an intrusion at or near a facility’s boundary. It can include fence sensors, buried detection systems, microwave sensors, LiDAR, and video analytics. Effective protection also requires visual assessment and a response plan so operators can act on the information.
Layered perimeter security combines technologies with complementary capabilities and limitations. A fence sensor may detect climbing or cutting, another sensor may monitor movement through a protected area, and cameras can help operators assess the event. The layers should reflect the site’s threats and give each technology a clear role in detection, assessment, or response.
In the manufacturer terminology Jeff Drews describes, a false alarm is generated without an external stimulus, while a nuisance alarm is triggered by activity that does not represent the intended intrusion threat. Wildlife, vegetation, and weather-related movement can cause nuisance alarms. A low published false alarm rate alone does not establish how many unwanted alarms operators will receive.
In a distributed acoustic sensing, or DAS, system, fiber-optic cable acts as a sensing element that detects disturbances along its length. Signal processing can help locate activity and distinguish patterns associated with events such as digging, fence climbing, or cutting. Performance depends on the system, the installation, and configuration for the site’s conditions.
A 2D LiDAR monitors a sensing plane, while a 3D LiDAR measures activity within a volume. Three-dimensional data can support more flexible detection zones and target assessment. The 3D system Jeff demonstrated also allowed exclusion zones. Detection range, classification, and performance in adverse weather remain dependent on the product, configuration, and application.
Sensor fusion combines information from multiple sensors to help interpret an event. For example, microwave detection, LiDAR location data, and camera imagery can contribute different information about the same activity. Its effect on security depends on how the system combines those inputs and whether a valid detection from one sensor can still generate an alarm.
AND logic requires both selected sensors to detect an event before generating an alarm. OR logic allows either sensor to trigger an alarm. AND logic can reduce unwanted alarms, but it can also withhold a valid detection when the second sensor does not confirm it. For the high security applications he discussed, Jeff emphasized preserving independent alarms while using combined information to improve assessment.
An alarm rule that requires a particular classification can miss an intrusion if the system does not classify the target in time. For example, a person may cross a narrow detection zone before the system identifies them as human. Zone dimensions, target speed, environmental conditions, and alarm rules should be tested against the movements the facility needs to detect.
Test both intrusion detection and nuisance alarm performance under conditions relevant to the facility. Include the required movements, such as walking, running, or crawling, along with relevant weather and normal site activity. Confirm what operators receive when only one sensor detects an event and whether camera coverage and response procedures support timely assessment and action.
Some compatible upgrades can reuse existing wiring, power supplies, poles, and mounting hardware while replacing sensor heads. Drews noted that this depends on the equipment and the upgrade being considered. New processing or detection capabilities are not automatically available through a firmware update, so compatibility and installation requirements should be confirmed for the specific models.