SECURITY SYSTEM DESIGN ANALYSIS

BACKGROUND OF THE INVENTION

The present invention relates to the field of security systems. In particular, the present invention relates to analyzing the design of an intrusion security system.

An intrusion security system detects specific events at a building or asset, typically with individual sensors that respond to security or safety breaches. The types of sensors used in an intrusion security system depend on various factors, including the type and location of assets to be protected, the physical layout and design of the site, environmental characteristics of the site, and usage patterns of site occupants. The most common types of sensors used in residential, commercial, and financial applications are passive-infrared (PIR) motion detectors, door contact sensors, glass break sensors, and seismic (i.e., vibration) sensors.

When a sensor is triggered, an alarm signal is sent to a call center where the data is logged and an operator is informed. The operator then either determines that the alarm is a false alarm (i.e., caused by something other than an intruder, fire, flood, or monitored machinery failure), or calls an appropriate agency (such as a guard or the police) to resolve the problem. The security system is designed to detect single premise events. However, increased detection rates come at the cost of higher false alarm and false dispatch rates. A false alarm occurs when a security system detects alarm status erroneously as a result of events such as user error, environmental triggering of sensors, or equipment failure. A false dispatch occurs when the call center, after being unable to verify the cause of an alarm by calling the premises or a property contact person, notifies a responding authority that visits the premises and finds no evidence of a threat to the premises.

False alarms and false dispatches introduce significant overhead into the security market. In addition, false dispatches compromise the

level of security provided to the end user of the security system, since resources that could be dedicated to responding to legitimate alarms must instead be used in responding to the false dispatches. Consequently, security systems must be designed and operated such that intruders are detected while innocuous events, such as authorized entries and benign environmental conditions, are ignored. However, security systems are often designed based only on past experience and informal rules of thumb. The system configurations and technologies designed and installed at security system sites are thus subject to human error and miscalculation. This can result in the misapplication of sensors and other security system equipment, resulting in security system designs susceptible to the occurrence of missed detections and false alarms.

BRIEF SUMMARY OF THE INVENTION

The subject invention is directed to analyzing the design of an intrusion security system for a location. A virtual spatial representation of the location is first generated. Virtual security system components, which are representative of actual components in the intrusion security system, are then integrated into the virtual spatial representation. The intrusion security system is then analyzed with regard to probability of intruder detection and probability of false alarm detection based on placement of the virtual security system components relative to the virtual spatial representation.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a system for generating a security system design according to the present invention.

FIG. 2 is a flow diagram for a process of designing and analyzing a security system according to the present invention.

FIG. 3 is an example of a virtual spatial representation of a structure including integrated environmental elements and sensors. DETAILED DESCRIPTION

FIG. 1 is a block diagram of a system 10 for generating a security system design according to the present invention. System 10 -includes

display 12, processor 14, and input device 16. Processor 14 includes security system components database 20, spatial layout application 22, controller 24, security system analyzer 26, and scenario database 28. Display 12, processor 14, and input device 16 may be provided in a single integrated unit such as a personal computer.

Input device 16 provides input signals to processor 14 that are received by controller 24. Input device 16 may be any device capable of receiving commands from a user and providing a signal related to the commands to controller 24, such as a keyboard, mouse, electronic drawing tablet, and the like. Security system components database 20 provides an input to controller 24 and includes data that may be stored on a storage medium such as a hard drive. Spatial layout application 22 and security system analyzer 26 may be computer programs or applications executed by controller 24. Scenario database 28 provides an input to security system analyzer 26 and includes data that may also be stored on a storage medium such as a hard drive. Security system components database 20 and scenario database 28 may be stored on the same storage medium or different storage media. Controller 24 is a microprocessor based controller that processes information from the various components of processor 14 and provides signals to display 12 to display the processed information in viewable form.

System 10 is used to generate, model, and fine-tune security system designs in a virtual environment. FIG. 2 is a flow diagram for a process of designing and analyzing a security system using system 10. A virtual spatial representation of the location at which the security system is to be installed is first generated (step 30). In system 10, spatial layout application 22 is used to create a virtual spatial layout of the location in which the security system is to be installed. A pre-existing virtual spatial layout of the location may also be loaded into system 10. In one embodiment, spatial layout application 22 is a computer-aided design

(CAD) program. Spatial layout application 22 may be presented on display 12 via a graphical user interface (GUI) to simplify the process of

using spatial layout application 22.

Spatial layout application 22 is used to first define the basic interior and exterior layout (e.g., walls, support structures, etc.) of the location in which the security system is going to be installed. When the basic layout of the structure has been defined, various features of the location must be added to the basic layout that are relevant to evaluating the performance of a proposed security system design. These features include points of entry, environmental features capable of triggering a false alarm in the security system, and asset locations and values. In order to analyze vulnerabilities and intruder detection probabilities of a location, portions of the location boundary where intruders can enter or exit the location must be defined on the virtual spatial layout. These points of entry include exterior doors and windows of the location. Doors and windows are characterized by their location on the virtual spatial layout and their intrusion probabilities. In one embodiment, the variable probabilities of intrusion through doors and windows are specified as a discrete set of values that equate to intrusion probabilities ranging from 0.1 (very low) to 0.9 (very high).

When the structural layout of the location has been completed, spatial layout application 22 is then used to define environmental elements at the location that are capable of triggering security sensors after the security system is activated. For example, sources of heat that could trigger infrared sensors, such as heating vents or electrical equipment, are defined on the virtual spatial representation. Also, sources of motion that could trigger a motion sensor, such as a banner hanging from the ceiling moved by the location circulation system, are defined on the virtual spatial layout.

The positions, footprints, and values of assets to be protected by the security system are then defined on the virtual spatial layout. The assets may take the form of point assets and region assets on the virtual spatial layout. Point assets cover a generally circular region on the virtual spatial layout, while region assets cover a generally rectangular region on

the virtual spatial layout. In both cases, the total value of the asset is distributed generally uniformly over the defined asset footprint. For example, if a safe or other secure asset storage facility is to be protected, a region asset having the general shape of the safe is defined on the virtual spatial layout, and the value of assets contained in the safe is assigned to the region asset. Thus, for a total asset value V ₀ over an asset region R having a footprint area A _R , the distribution of asset value (in units of value per square foot) can be expressed as:

V(r) = (Equation 1). In general, IR is one within the region R and zero everywhere else:

⁷ * ^{(F) s} ln ^; -! » (Equation 2).

When the virtual spatial layout of the location has been defined using spatial layout application 22, virtual security system components from security system components database 20 are then integrated into the virtual spatial layout (step 32). Security system components database

20 stores virtual models of actual security system components, such as sensors operable to detect intruders and ancillary control devices (e.g., panels and keypads) for controlling operation of the security system. The virtual sensor models in security system components database 20 may include, but are not limited to, models for such sensors as passive infrared (PIR) motion sensors, 360° motion sensors, door contact sensors, window contact sensors, glass break sensors, and dual technology sensors (described below). Each virtual model includes variable parameters related to the functionality of the actual security system component that it models. These parameters are configurable for a particular security system design (step 34).

PIR motion sensors and 360° motion sensors detect passive infrared (IR) emissions within a domain. PlR motions sensors detect passive IR emissions within a pie-shaped domain, while 360° motion

sensors detect passive IR emissions within a circular shaped domain. PIR motion sensors and 360° motion sensors are triggered when a source of IR emissions (e.g., an intruder) moves across the field of view of a sensor (i.e., within the detection domain). The PIR motion sensor model stored in security system components database 20 may include variable parameters such as the sensor orientation angle, directional detection range, angular detection range, sensor position, and maximum probability of intruder detection. The 360° motion sensor model from security system components database 20 may include variable parameters such as directional detection range, sensor position, and maximum probability of intruder detection (with a fixed angular detection range of 360°). For both PIR motion sensors and 360° motion sensors, the probability of detection P _d of an intruder at position F for a motion sensor at position F ₀ having a range R, a maximum probability of detection P _rf ^max , and a general parameter A (based on the angular and directional range of detection) may be given by:

^p « ( ^? ) (Equation 3).

Door contact sensors detect the opening of a door through the interruption of a magnetic circuit attached to the door and its frame. Similarly, window contact sensors detect the opening of a window through the interruption of a magnetic circuit attached to the window and its frame. The door and window contact sensor models stored in security system components database 20 may include variable parameters such as sensor position and probability that the sensor will detect an intruder passing through the door or window on which the sensor is mounted (i.e., the sensitivity of the sensor). In other words, door and window contact sensor models in security system components database 20 are characterized by their reliability specified in terms of their probability of detection Pφ

Glass break sensors detect the acoustic energy emitted when a window pane or other glass is broken. Glass break sensors typically have a half-circle shaped detection domain. The glass break sensor models stored in security system components database 20 may include variable parameters such as sensor position, directional range of detection, sensor orientation, and the maximum probability that the sensor will detect an intruder breaking a window pane. The probability of detection P _d of a glass break sensor at position r may be given by Equation 3.

Dual technology sensors are, for example, a combination of PIR and microwave motion sensors. As described above, PIR motion sensors are passive recipients of IR energy emitted by intruders. Microwave sensors are active emitters of microwave radiation that is reflected off of intruders and detected by the sensor. The combination of passive IR and active microwave detection allows for greater discrimination between intruders and false alarms. The dual technology sensor models stored in security system components database 20 include similar customizable parameters to those of the PIR and 360° motion sensors. The probability of detection P _d of a dual technology sensor at position r may be given by Equation 3. After the security system has been designed and laid out on the virtual spatial representation, security system analyzer 26 is used to analyze the security system design (step 36). Security system analyzer 26 combines the variable parameter information from each of the virtual security system components with information from scenario database 28 to analyze the security system design. Scenario database 28 includes data related to general and location-specific intruder threat scenarios and false alarm scenarios. More particularly, the general scenarios include simulation information about common intruder and false alarm causes, and the location-specific scenarios include simulation information particular to the security system location (e.g., likely paths of movement by an intruder through the location).

Security system analyzer 26 performs static and dynamic analyses

on the security system design. The static analysis includes computations of a static probability of detection by a set of the sensors, an asset risk analysis based on the static probability of detection and asset values, and a static probability of detecting various types of false alarms. The dynamic analysis includes computations of probabilities of detecting intruders along shortest and stealthiest (i.e., most avoiding of sensor detection) paths from points of entry to asset locations, and a vulnerability map of the security system location based on the shortest and stealthiest intruder paths. For both static and dynamic analyses, various probability calculations are made based on the layout of the security system. The results of these probability calculations may then be superimposed on the virtual spatial layout in the form of a graded map to indicate areas of high and low intruder detection probability, false alarm probability, asset risk, and shortest and stealthiest path location vulnerability. The following discussion sets forth the methodology used for the probability calculations.

A static probability of detection may be calculated for motion sensors and glass break sensors in the security system. For an individual motion sensor in the security system design, the probability of not detecting an intruder at position F is given by 1 - P _d (r ), and the probability that all motion sensors in a set of N _mo ti _o n motion sensors will miss the intruder is given by:

ll(l -P _dtk (r)) (Equation 4).

Ar=I

Thus, the net probability that an intruder at position F will be detected by at least one motion sensor is given by:

5).

Similarly, the net probability that an intruder at position F will be detected by one of a set of N ₉ b glass break sensors is given by:

^ ^8b( ^ ⁾ = ^l -π( ¹ - ^p ^ ^(?) ) ( ^E q ^uat| on 6).

The results of these probability analyses may then be superimposed on the virtual spatial layout for all positions 7 at the location in the form of a graded or color coded map that includes different colors for different probabilities of detection.

An asset risk analysis based on the static probability of detection and asset values may also be calculated and displayed as a risk map. A risk map displays the asset value weighted probability that motion sensors in asset-occupied areas will not detect an intruder. As discussed above, the probability that no motion sensor will detect an intruder at location 7 is given by I-Pj ¹¹⁰ " ⁰ "^) . If asset k occupies domain D(k), the risk map for a premise is given by:

R ⁽ 7)≡ J(l-Pr ^tion (^))-^)(?)-(l-^ ^motion (?))-^('") (Equation 7)

where I _D(k) (P) is the indicator function over domain D(k), N _asS ets is the number of assets, and D is the union of the asset domains.

D (Equation 8).

Using the above foundation for probability of detection, a false alarm detection table may be created. A false alarm detection table shows detection probabilities for environmental sources of false alarms. As described above, these may include non-intruder sources of motion and heat. The entry in row a and column s represents the false alarm probability (F _a , _s ) of detecting the environmental source a with motion sensor s. Thus, each entry in the table is given by:

F _α,s = P _d T ^α (r _α ) (Equation 9) where f _α is the position of an environmental element that is a potential false alarm source, and P™ ^&on (r) is the probability that motion sensor s will detect an intruder at location 7.

Dynamic probability analyses calculate a shortest path detection probability, a stealthy path detection probability, and an overall dynamic vulnerability of the location. Dynamic detection probability analyses calculate the probabilities of detecting an intruder moving along shortest and stealthiest paths from points of entry into the security system premises to assets located in the interior of the premise. The former paths model intruders trying to retrieve an asset in as little time as possible without regard to whether their motions trigger the alarm system. The latter paths model intruders trying to retrieve an asset without triggering an alarm system no matter how long it takes to reach an asset in a premise. A dynamic vulnerability analysis calculates the probabilities of detecting an intruder moving along shortest or stealthiest paths from points of entry into a premise to a sample of locations in the interior of the premise that may or may not contain an asset. This analysis is used to analyze the vulnerability of a site for which asset information may not be available or where the assets are uniformly distributed.

As discussed above, the probability of not detecting an intruder at r with motion sensor s is given by 1 -P/(r) . For a set of N _s sensors, the detection probability is:

P,(?) = l-π(l-P;(r)) (Equation 10).

This equation can be applied to path positions and multiple motion sensors. Each motion sensor is modeled as being able to detect an intruder only along a discrete set of uniformly-spaced rays or curtains / that originate at the sensor's location and extend a distance R from it. PIR motion sensor curtains occupy a pie-shaped wedge region, and 360° motion sensor curtains occupy a circular region. If A _sj represents the rth curtain associated with sensor s, and N _s is the number of motion sensors protecting the location, the probability of detecting an intruder moving along path r is given by:

Ns N _c (s) iT ^fal Cr) = l-rin( ¹ - ⁱ ?CrnA _M )) (Equation 11) s=l λ=\ and

PS(φ) = O (Equation 12). The number of curtains associated with a motion sensor is defined to be:

N _c (s) = ^2'5RsA<Ps (Equation 13)

^s where Aφ _s is the angular separation between neighboring curtains, l _s is the distance between the endpoints of neighboring curtains, and R ₃ is the sensor's detection radius.

The shortest path detection probability analysis calculates the probabilities of detecting an intruder that is moving through the security system location using the shortest path between each point of entry and each asset defined on the virtual spatial layout. If r _e ^sh _a ^ort is the shortest path from entry point e (e.g., a window or a door) to asset a, then the shortest path between entry point e and asset a is given by:

rf _α ^ort ≡ argmin \ds ; ds ² = dx ² + dy ¹ (Equation 14).

A shortest path detection probability table may also be created wherein an entry P* ^h0It (e,ά) in row e and column a of a shortest path detection probability table is given by:

Pt ^ort (e,ά) a P _d ^moύm (r;i ^ort ) (Equation 15). The stealthiest path detection probability analysis calculates the probabilities of detecting an intruder that is moving through the security system location using the stealthiest path between each entry point and each asset defined on the virtual spatial layout. The stealthiest path between two points is the path having the least probability of detection by the security system design. If r _e ^s ^ ^althy is the stealthiest path from entry point θ (e.g., a window or a door) to asset a, then the stealthiest path between entry point e and asset a is given by:

rr ^lthy _≡ argmm(pr ^ti ° ⁿ (r _e , _fl )) (Equation 16)

T _e ,a where P™ ^ύoa (r _e J is given by Equation 11 above. A stealthiest path detection probability table may also be created wherein an entry P/ ^ealthy O,α) in row e and column a of a stealthiest path detection probability table is given by:

P; ^tealthy (e,a) ^ P7 ^tion (r _e ^s 7 ^lthy ) (Equation 17)

Dynamic vulnerability calculations identify the most vulnerable parts of a protected site based on shortest and stealthiest paths from points of entry. A vulnerability map based on the shortest paths from points of entry depicts the probability of detecting an intruder reaching every point on a map, assuming that the intruder always takes the shortest path from any point of entry to each point. The vulnerability map takes into account all sensors, all points of entry, and the probability of entry from each point of entry. If a security system location has N _p points of entry, the normalized probability p(k) of an intruder entering the location via the /rth point of entry is given by:

/W (Equation 18),

and

Y ₁ P(Ii) =I (Equation 19),

Jt=I where P _a (k) is the variable intrusion probability for point of entry k as set during generation of the virtual spatial layout. The virtual spatial layout is then partitioned into a plurality of non-overlapping subregions. For each subregiony, the vulnerability of the subregion is given by:

P _d U) = ∑p(k)-PAKD (Equation 20),

A-=l

where T^° ^A is the shortest path from the location of the /cth point of entry to the center of the yth subregion, and P _d (T^ ^A ) is the probability of detecting an intruder along the shortest path from point of entry k to subregion/ A vulnerability map based on the stealthiest paths from points of entry depicts the probability of detecting an intruder reaching every point on a map, assuming that the intruder always takes the stealthiest path from any point of entry to that point. As before, the vulnerability map takes into account all sensors, all points of entry, and the probability of entry from each point of entry. If a security system location has N _p points of entry, the normalized probability p(k) of an intruder entering the location via the /cth point of entry is given by Equations 18 and 19. If the virtual spatial layout is then partitioned into a plurality of non-overlapping subregions, for each subregion j, the vulnerability of the subregion is given by:

P _d U) = ∑Pik) ^• P _d (T***) (Equation 21 ),

where r^ ^althy is the stealthiest path from the location of the /cth point of entry to the center of the yth subregion, and i ^> _d (r^ ^althy ) is the probability of detecting an intruder along the stealthiest path from point of entry k to subregion j.

When the security system design has been analyzed pursuant to the above methodology, the results of the analysis are reviewed to determine whether the security system design is satisfactory in terms of the functional and cost goals of the security system (step 38). If the results of the analysis are deemed satisfactory, the security system design process is ended (step 40), and the design may then be implemented at the location.

If the results of the analysis are deemed unsatisfactory, properties of the security system design may then be adjusted to achieve desired

functional and cost goals of the security system (step 42). For example, the variable parameters of the virtual security system components may be adjusted to provide the desired functional results. In addition, the types and amount of security system components used may be adjusted to achieve the desired functional and cost goals. Furthermore, the positioning of the virtual security system components may be adjusted, either to increase the probability of intruder detection by the security system, or to reduce the occurrence of false alarms in the security system. When adjustment of the security system design has been completed, the security system design is then analyzed using the above methodology to determine if further adjustments need to be made (step 36). Thus, because security systems are designed and revised based on quantitative analysis of virtual representations of actual system components, design flaws due to human error and miscalculation are reduced, thereby improving intruder detection while reducing false alarm detection.

FIG. 3 is an example of a virtual spatial representation 50 of a location as designed during steps 30 and 32 described above. Virtual spatial representation 50 includes outer boundary 52, asset footprint 54, doors D1 and D2, windows W1 and W2, point of entry (i.e., door and window) contacts C1 , C2, C3, and C4, glass break sensor GB1 , and PIR sensors PIR1, PIR2, and PIR3. PIR2 and PIR3 are mounted on an interior wall, and asset 54 is located adjacent to an interior structure. Virtual spatial representation also includes environmental elements capable of triggering a false alarm. These elements, shown as circled letters, include a banner B and two heat sources H. When virtual spatial representation 50 has been completed, the virtual security system components are configured to have the desired properties based on the functional goals of the system (as described with regard to step 34 above). The security system design is then analyzed with regard to probability of intruder detection and probability of false alarm detection (pursuant to the methodology described with regard to step 36 above).

Finally, the analysis is reviewed and the security system design is adjusted as necessary to achieve the desired functional and cost goals of the security system (as described with regard to steps 38-42 above).

In summary, the present invention is directed to systems and methods for designing and modeling an intrusion security system for a location. A virtual spatial representation of the location is first generated. Virtual security system components, which are representative of actual components in the intrusion security system, are then integrated into the virtual spatial representation. The intrusion security system is then analyzed with regard to probability of intruder detection and probability of false alarm detection based on placement of the virtual security system components relative to the virtual spatial representation. The virtual security system components are stored in a database and may be configured based on actual functional properties in the security system. After the intrusion security system is analyzed, the virtual security system components may be repositioned as necessary to assure that intruders are detected while false alarms are avoided.

Although the present invention has been described with reference to examples and preferred embodiments, workers skilled in the art will recognize that changes may be made in form and detail without departing from the spirit and scope of the invention.