(Continued from Part 1.)
The final aspects you need to consider when planning a security system are the types of threats you need to be able to detect. If you live in a wooded area where there are a lot of experienced woodsmen and hunters, you’ll have to consider how to detect people that know how to move silently and effectively and are more likely to notice things like tripwires or trail cameras. On the other hand, if you’re in an area that may primarily experience urban sheeple migrating in search of resources after a disaster, your security situation will be simplified, since most of those folks will take the path of least resistance with minimal regard for stealth. An absolute worst-case scenario would be trying to set up a security system to detect people that have military training and equipment, since they will tend to be more observant about wires, alarms, infrared light, etc. As with any system, the more complex and extensive you make it, the harder it will be to maintain and the more likely you are to get a lot of false alarms.
Once you understand the primary drivers for your security system, you should begin to develop an overall approach to how you want to implement it. The best approach is to leverage what the military refers to as ‘defense in depth’ – layers of detection capabilities at different distances from your house or property. The goal is to have multiple opportunities to detect an intruder in case they get by one of the perimeters. Remember that to stay safe in a fixed location you need to be able to detect every intruder approaching from any possible direction at any time, but an intruder only has to be lucky or good enough to get through your detection perimeter once to do some harm.
Reducing Your Attack Surface
In the cybersecurity industry there’s a concept known as an ‘attack surface’, which is the sum of all of the possible ways a computer can be attacked. To minimize the resources required to monitor all possible paths, one common recommendation is to block off as many attack vectors as possible so they can’t be used. This concept can also be applied to designing a physical detection system by minimizing the number of possible ways an intruder can approach and move through your property, thereby reducing the number of paths you need to monitor.
There are a lot of options for accomplishing this, such as adding obstacles or landscaping. For example, a high fence that completely surrounds your property would force an intruder to either approach through a single gate or attempt to scale the fence, where they can be more easily detected and dealt with. However, for many of us this is not an option, so you should focus on limiting possible approaches as much as possible. Even if you can’t fully enclose your property with fencing you may be able to install some stretches to funnel an intruder down certain paths, making them easier to detect. Rolls of concertina (razor) wire can also be used, or thorny bushes such as Honey Locust, Blackberry/Raspberry, Roses, Bougainvilleas and Pyracantha can be planted. You can even pile up bunches of loose branches and other debris, which most people will naturally go around. If you live near a stream or other body of water you may be able to divert some water to low-lying areas and turn them into mud pits, which most people will avoid.
If you’re trying to protect a building you should also consider radial obstacles – these are fencing, concertina wire, thorn bushes, etc. that start at the building and move straight out. If you have a house where an attacker could get close and move around the house unseen below the window level, radial obstacles would force them to move away from the building and potentially expose them to observation and detection.
When planning obstacles to funnel intruders you should attempt to reduce the potential for those obstacles to provide the intruders with cover and concealment. While a solid-wood fence may force an attacker to move down a different path, it is also something they can potentially hide behind, so chain-link fencing might be a better option. When planning obstacles to funnel intruders, draw them out on a map and simulate them with full-sized wood or cardboard mock-ups to see how they impact your view of the approaches. If you’re forced to deal with obstacles that provide cover or concealment, you may need to plan on some additional capabilities that can improve your ability to detect and observe an intruder hiding behind those obstacles (more on that later).
Once you’ve defined your requirements and general approach, your next consideration should be what types of sensors and alerting you want to implement. Note that for the purposes of this article I’m not covering direct visual observation as a detection method, although it can be utilized as one in combination with the other types of sensors if you have the personnel available to keep watch. Activities such as active patrolling can also extend your awareness of your surroundings, but it requires qualified people.
The sensor (or trigger) is the component that actually detects the presence of a person or vehicle and triggers an alert, and comes in two general varieties:
- Direct contact – A person/vehicle directly interacts with the sensor, or something the sensors is connected to, to trigger an alert. These include tripwires, pressure mats, door opening sensors, etc.
- Proximity – The nearby presence of a person or vehicle causes the sensor to trigger an alert. These include motion detectors and magnetic vehicle sensors.
Direct contact sensors generally have the advantage of fewer false alerts, but they can only usually provide coverage for a relatively small area or limited range of intruder actions. Proximity sensors let you cover a broader area, but tend to be more prone to false alarms from animals, moving branches, blowing snow, etc. You can and should leverage both types of sensors, using each where it makes the most sense for your requirements. Consider a situation where you want to alert on someone entering a room that has several doors and windows – you could install one motion sensor to cover the doors and windows, which may be subject to false alarms from your dogs or sunlight, or you could wire sensors directly to each door and window.
Alerting is how your security sensors interact with humans and notify you when something happens, so you also need to consider how you want to be alerted if a sensor is triggered. There are two possible implementations to consider:
- Centralized/Remote – The remote sensors send an alert signal to one central location, like a security station.
- Local – Alerting occurs where the sensor is triggered.
A centralized system has the advantage of collecting your alerts in one location and can provide you with better spatial awareness of which sensor was triggered, but it can be more complex to implement since you’ll need some method to send the sensor signal to that central location (either wires or wirelessly). You’ll also need to have a person at or near the centralized alerting location to monitor for alerts, or have audible alarms loud enough to wake you up. Another significant advantage of centralized alerting is that since the audio or visual alarm doesn’t happen near the sensor, an intruder may not realize they’ve been detected, which could give you more time to intercept them and take action.
Local alerting means some form of sound or visual signal is generated at or near the sensor’s location, such as a tripwire ringing a bell, so it’s a lot easier to implement. One down side of local alerting is that with sound-based alerts (e.g. blanks, siren, bells, etc.) it may be difficult to localize exactly where the alert is coming from; another disadvantage is that the intruder will know they’ve been detected and may take rash action. Local visual alerts such as flares or smoke bombs also require that someone see them to be effective, and some of them may not be as visible at night or in heavy weather conditions such as rain or snow. One possible approach would be to combine audible and visual alert signals for local alert sensors so they are more likely to be noticed and localized across a wide range of conditions.
As with sensors, you can apply a combination of both centralized and local alerting to meet your requirements. For example, you could have sensors at distant strategic locations that that uses aerial flares combined with audio alarms to alert you that an intruder has entered your area, which would allow someone to get to the centralized alert console to track their progress and radio directions to a response team.
For my personal system I’ve planned and (partially) implemented a four-layered sensor approach – the more distant remote sensors at strategic locations will use local alerting with 12GA shotgun blanks and flares, and the outer, inner and interior perimeter sensors all feed back to a central console in my house with a mix of wired and wireless signals. Note that I’ve planned, designed and built the remote strategic blank/flare sensors, but I won’t install them until things really go pear-shaped. I’ve drawn a map of my house and property on a piece of ¼” plywood and installed 12V automotive panel lights/buzzers that light up and buzz when the corresponding sensor is triggered. Here’s an example of what the board could look like (NOT my actual property layout!):
I’ve run the wires to the wired sensor’s locations over the course of normal yard and house work. The sensors are a combination of tripwires, motion sensors and Reed switches that cover obvious trails in the woods, the area right around my house and doors and windows in my house. When any sensor is tripped, it closes the 12V circuit and triggers the corresponding light/buzzer on the console. I’ve also installed some switches on the console that allow me to disable any of the alert zones if necessary. The system is not that complicated, and only requires a basic knowledge of how to wire a light/buzzer and a switch to a 12V power supply.
(To be continued tomorrow, in Part 3.)