Object proximity/security adaptive event detection

专利摘要:
The security system incorporates security rules and procedures and reasoning systems that are designed to allow the state as carefully as possible. Two unique aspects of the system facilitate the enforcement of rules and procedures in a careful manner. First, a transponder that can be triggered and sensed from a distance is preferably used to identify both items and individuals. This remotely sensed identifier is processed by the inference system to determine whether each identified item is or is likely to be authenticated such that the identified individual moves or retrieves the identified item from a secure location. Second, the system continues to change and optimize rules and procedures based on the determination of security events. An initial set of rules is created for a security system, which generally prevents a security item from moving out of a secure location, except when a particular individual is authorized to move a regulated item out of a secure location. Thereafter, the security system is configured to enforce these security rules and procedures and to gather feedback from authorized security personnel regarding the effectiveness of the security rules and procedures enforced. A learning system is coupled to the security system that is configured to change existing rules or generate new rules in response to comments from authorized security personnel. By dynamically adjusting the security rules and processes, the breach of the security system against the monitored individual is substantially reduced, and the system is optimized based on continued opinion.
公开号:KR20020029382A
申请号:KR1020027002068
申请日:2001-06-15
公开日:2002-04-18
发明作者:케이쓰 에. 마티아스；야메스 데. 쉐퍼
申请人:요트.게.아. 롤페즈；코닌클리케 필립스 일렉트로닉스 엔.브이.；
IPC主号:

专利说明:

Target proximity / security adaptive event detection {OBJECT PROXIMITY / SECURITY ADAPTIVE EVENT DETECTION}
[2] Security systems are common in the prior art. As computer and database systems emerge, inventory security systems are also becoming widespread. The PCT patent application (WO 97/15031) entitled "Article Inventory Tracking and Control System" published on April 24, 1997, describes each item listed as a "marker." Describes a system that is uniquely identified through. The user associated with the security facility is also uniquely identifiable, for example, via an identification card with a magnetic strip comprising a unique identifier. The user places the items listed with the user's identification in a "check-out / check-in" device. If the user is authorized to remove the device from the security facility, the "marker" is switched to an inactive state. In a retail environment, the user is authorized to move the device after a debit has been registered in an account associated with the user's identification, such as the user's credit card account. Each exit from the security facility includes a sensor for active markers. If the marker of the listed item is not deactivated by the check-out / check-in device, the sensor will detect the active marker, and an alarm event is triggered to prevent movement of unauthorized items. triggered). In a similar manner, a user can return a list item to a security facility by providing the item to a check-out / check-in device. When the list item is checked in, the device reactivates the item's marker and updates the database file to reflect that the user returned the list item. Existing applications of the system include automated check-out / check-in processes for lending libraries, video rental stores, and the like. U.S. Patent No. 4,881,061, entitled "ARTICLE REMOVAL CONTROL SYSTEM," issued November 14, 1989, operates similarly.
[3] U.S. Patent No. 5,886,634, entitled "ITEM REMOVAL SYSTEM AND METHOD," issued March 23, 1999, which is incorporated by reference herein, is incorporated by reference in its entirety. It provides a low intrusive system using wireless-ID tags. The database associates each identified item with one or more persons authorized to move the item. When an item is detected at the exit without an authorized person, a warning sounds. The system also interfaces with the list control system and can provide the aforementioned capabilities, such as automated check-in and check-out systems.
[4] In prior art systems, a database of certificates for each secured item in the list should be kept up to date. Because of the overhead typically associated with maintaining an inventory security system, the rules and procedures enforced are relatively static and simple. Such a system can be well suited for book rental or retail environments and is convenient compared to conventional manned check-out stations, but such a system is unlikely to withstand a normally unsecured environment.
[5] In the environment of an office or a laboratory, for example, even if the collection is stolen in such an environment, the employee is generally not subject to security. This lack of security may be based on reluctance to prove lack of trust in the employee; May be based on logistical difficulties, such as exit queues, caused by requiring each employee to check out a list item each time the item is moved out of a secure facility; It can be based on anticipated annoyances that false alarms can trigger. Similarly, in many large organizations, or large facilities, attempts to map each identified item in the facility with a group of individuals authorized to move the item may be impractical.
[1] TECHNICAL FIELD The present invention relates to the field of security systems, and in particular, to security systems that adaptively generate and change security rules and parameters based on conventional events.
[9] 1 is a block diagram of an example of a security system in accordance with the present invention;
[10] 2 is a flow chart of an example of a security system in accordance with the present invention.
[11] 3 is a block diagram of an example of a learning system for use in a security system in accordance with the present invention.
[12] 4 is a flow diagram of an example for updating a rule set of a security system established in accordance with the present invention.
[6] It is an object of the present invention to facilitate the task of automating a security system. It is a further object of the present invention to minimize the breach of security processes for monitored individuals. A further object of the present invention is to facilitate the dynamic modification of the security process invoked by the security system.
[7] These and other objects are achieved by providing a security system that incorporates security rules, procedures, and reasoning systems that are designed as carefully as the situation permits. Two unique aspects of the system facilitate the enforcement of rules and procedures in a careful manner. First, a transponder that can be triggered and sensed from a distance is preferably used to identify both items and individuals. This remotely sensed identifier is processed by the inference system to determine whether each identified item is or is likely to be authenticated such that the identified individual moves or retrieves the identified item from a secure location. Second, the system continues to change and optimize rules and procedures based on the determination of security events. An initial set of rules is created for a security system, which generally prevents a security item from moving out of a secure location, except when a particular individual is authorized to move a regulated item out of a secure location. The security system is then configured to enforce these security rules and procedures and to receive feedback from authorized security personnel regarding the effectiveness of the security rules and procedures enforced. A learning system is coupled to the security system that is configured to change existing rules or generate new rules in response to comments from authorized security personnel. By dynamically adjusting security rules and procedures, the breach of security systems for monitored individuals is substantially reduced, and the systems continue to be optimized based on comments.
[8] The invention is explained more specifically by way of example with reference to the accompanying drawings.
[13] 1 shows a block diagram of an example of a security system 100 in accordance with the present invention. In a preferred embodiment, a transponder (not shown) is attached to the inventory item 102, such as a portable computer system, part of an office or lab instrument, and the like. Each exit exiting the secure location includes an area monitored by the item detector 120. Consistent with conventional transponder technology, detector 120 emits a trigger signal in the vicinity of the surveillance area. Detector 120 also detects emission from the transponder triggered by the trigger signal of the detector. Each transponder emits a unique code, which is associated with the inventory item to which it is attached. The unique code from the transponder is provided to the inference system 150 via the detector 120.
[14] In a preferred embodiment, another transponder (not shown) is attached to the individual 101 as a transponder, generally mounted on a security badge. Personal detector 110 examines the surveillance area and detects emissions from the transponder, similar to item detector 120 to determine a unique code associated with person 101. The unique code from the transponder is provided to the inference system 150 via the detector 110.
[15] Note that unique detectors 110 and 120 are illustrated for ease of understanding. A single detector system can be used to detect transponders associated with items or individuals. Any number of conventional collision avoidance techniques may be used to avoid interference, or “collisions” in both transponders, or in response from multiple transponders associated with multiple items 102. The transponder can be configured to be triggered by a different trigger signal. The item transponder may be triggered in one portion of the surveillance area, or in one time period, and the personal transponder may be triggered in another region, or another time period. Alternatively, all transponders may be triggerable by the same trigger. In one such embodiment, each transponder, or each class of transponders, may be configured to transmit on a different frequency. Each transponder may be configured to 'listen' the response of another transponder before initializing itself. Each transponder, or class of transponders, may be configured to transmit with a delay time different from the time the trigger signal is received from the detectors 110, 120. Each transponder, or class of transponders, can transmit using different CDMA code patterns and the like. Such techniques for distinguishing transmissions in a multi-transmitter environment, and combinations of these techniques, are common in the prior art.
[16] Other item and personal detection techniques may also be used. For example, an individual may be recognized through machine vision systems, biometric recognition systems, and the like. In a similar manner, a computer device can be programmed to periodically transmit a beacon signal, which can be used to identify a computer item or trigger another security subsystem.
[17] In general, the system 100 may provide one or more item identifiers to the inference system 150 via the detector 120 and at most one personal identifier to the inference system 150 through the detector 110. It is composed. Alternatively, when there are a large number of people in the surveillance area, localized detectors 110, 120 or direction-finding / place-determining detectors 110, 120 associate each person with the detected item. Used. Given the environment in which many items that multiple people require to carry are commonly encountered, the system 100 may be configured to provide each item identifier to multiple personal identifiers as needed. For ease of understanding, the present invention is provided later on the assumption that each detected item identifier is provided to reasoning system 150 with at most one personal identifier. In addition, the system 100 is preferably configured to distinguish the movement and return of the item at the security facility to facilitate subsequent processing. The divided surveillance area may be provided with an inlet and an outlet, for example, or direction-determination detectors 110 and 120 may be used. Alternatively, the system initially sets a flag that indicates that the item is in the secure area and is associated with each list item, and then each angle of the item in the entry / exit area representing each move / return. It can be configured to toggle the flag with subsequent detection.
[18] In a preferred embodiment, the inference system 150 processes the received item identifier and the personal identifier based on a set of security rules 145 as shown by the example flow diagram of FIG. 2. As shown by the continuous loops 210-260 in FIG. 2, the example inference system 150 of FIG. 1 continues to process item identifiers received from the item detector 120 of FIG. 1. Upon receiving the item identifier in step 210, in step 215 the inference system determines whether any security rules (145 in FIG. 1) apply to the identified item. For example, some items, such as samples, may be identified for listing purposes rather than for security purposes, and someone will be allowed to move such items from a secure location. If the security rule is applied at step 215, then at step 220 a personal identifier is received if so. As mentioned above, it is preferred that a transducer is provided as part of the security badge. If the person carrying the identified item (102 in FIG. 1) (101 in FIG. 1) has such a badge, an identifier of the person is received in step 220. If the person does not have a transponder, a null identifier is produced.
[19] The security rule 145 includes a rule associated with each identified item as one of item-specific rules, item-class rules, general rules, and the like. For example, a general rule may be: "Send alert A if any item identifier is received without personal identifier"; Or "If any item identifier is received between midnight and 5 am, and a personal identifier is not X, Y, or Z, a warning B is issued." For example, the item-class rule may be: "Send alert C if any lab-class item identifier is received and no personal identifier is included in the lab list"; Or, if the cost associated with the item identifier is greater than $ 500 and the personal identifier is less than X, issue a warning D. For example, a particular rule may include: "If the item identifier X is received and the personal identifier is not Y, send a warning E"; Or, if an item identifier Z is received and a personal identifier does not exist in group A, a warning F is sent ". As will be apparent to one skilled in the art, the rules may also include "else" clauses, "case" clauses, and the like, which clauses may correspond to the correspondence between the identified item and the identified individual, or Further limit the security actions to be taken due to the lack of responsiveness.
[20] The term "warning" is used herein to include the results of a security assessment. These warnings include sounding audible alarms, sealing exit points from security facilities, turning on video cameras, calling remote security sites, and relocating to selected addresses. Sending mail, and the like. In a typical embodiment for an office or laboratory environment, an alert is generally a message on the display console for potential later action by security personnel to avoid bad consequences of false alarms, or excessive reaction to minor inconsistencies. Will include displaying. In some installations, authorized movement of the identified item may also trigger an alert, for example, notifying security personnel of "OK to remove". Note also that the principles of the present invention are not limited to security systems. The terms "security system", "warning" and the like are used to facilitate understanding. For example, the system 100 may be used in a field-service facility that builds a finite list of specific parts of a test device, and person X returns, "If someone returns an item identifier corresponding to an oscilloscope-type item, , Send a warning to X ". In this manner, system 100 can be used in conjunction with another system, such as a message system, and the rule is, "If the item identifier is X and the personal identifier is Y, then any message in the message system for person Y is sent. Send to X device ". Similarly, the surveillance area may include an audio output device, and the rule may be, "If the personal identifier is Y, say 'John, call me bill before you leave'", or "if the personal identifier is Y, Play message Y1 ". Such and other applications of system 100 with remote item and personal sensing capabilities will be apparent to those skilled in the art in view of this disclosure. Note that the structure of "if ..." is provided for ease of understanding of the above exemplary rule. As is common in the prior art, various techniques such as neural networks, fuzzy logic systems, transaction systems, associative memory systems, expert systems, etc., are used to achieve selection based on multiple inputs. Used.
[21] The security rules may be based on background or environmental factors such as the day of the week, the time of day, security status at the facility, and the like. The security state may include, for example, whether an alarm sounds, whether the alarm is a security or safety alarm, and the like. That is, for example, movement of any item and all items may be authenticated when a fire alarm sounds, while movement of a class of selected items may be blocked when an intrusion alarm sounds. If so configured, these environmental factors are provided by an environmental monitor (180 of FIG. 1) and received by the inference system (150 of FIG. 1) at block 230 of FIG. 2.
[22] If the security event is triggered by a combination of item identifier, personal identifier (if any), and environmental parameters (if any), then in step 240, an appropriate alert is issued. As will be discussed further below, comments based on the alert are collected at step 250, and these comments are used to update the security rules. After updating the rule in step 260, or if no security event is triggered in step 235, or if no rule is associated with the item identified in step 215, the process receives the next item identifier. To block 210 again. Optionally, in step 270, a log of the results caused by each received item identifier is maintained for later review and review by security or management personnel. According to another aspect of the present invention, the security system 100 of FIG. 1 includes a learning system 140, which changes the security rules 145 used by the inference system 150. It is configured to. The learning system 140 changes the security rule 145 based on the opinions converged in response to the alert via the security interface 130. The learning system 140 attempts to optimize the performance of the security system by reinforcing the correct behavior of the inference system 150 and preventing incorrect operation.
[23] In many large organizations or large organizations, attempts to map each identified item in the facility to a group of individuals authorized to move the item may be impractical. In such circumstances, the behavior of the security system will be in accordance with the organization's policies. In a non-automated environment, for example, some organizations will conduct mandatory investigations of all packages being transported in security facilities. Other organizations will conduct a "spot check" investigation of the package being moved. When either system is used first in an organization, inefficiency is common. As security staff gains experience, the system runs smoother. The particular person will be aware that the type of item for which the particular person is authorized to move normally is known and the like. Certain items are found to be particularly popular stolen items, such as computer accessories, while others are found to be popular mobile-and-return items, such as special purpose test devices and the like. It is recognized that most current security systems are not simple. The experience of a security staff relies on providing a reasonable and effective tradeoff between the need to maintain security and the inconvenience caused by the security system. In general, security resources consume more of the rarer events than regular ones, although evil robbers can take advantage of the weakened security of being concentrating on regular events.
[24] In accordance with this aspect of the present invention, the learning system 140 emulates the learning behavior of security personnel with an added advantage of knowing what items are being moved or brought from the facility. Using techniques common to the prior art, the learning system 140 may, for example, detect the reasoning system 150 based on the security personnel's judgment of the alert sent out from the reasoning system 150 via the security interface 130. We collect opinion from). When the security system 100 is first installed, for example, many warnings will be issued. The security officer may alert all or some warnings, such as asking the selected identified individual 101 for proof of authentication to move the item 102, or confirming the individual's supervisor about such authentication. Will take some action. Generally, such measures are known to carry security items 102 to individuals subject to such sampling investigations, thereby increasing the efficiency of such sampling investigations (whether or not the learning system is used). This is the same action taken by security personnel on non-automated systems.
[25] In accordance with this aspect of the present invention, to further improve the efficiency of the security operation, the security officer notifies the reasoning system 150 of the results of the sampling check. The inference system 150 processes this opinion in a form suitable for processing by the learning system 140. For example, reasoning system 150 may include a particular 'input stimuli' (personal identification, item identification, environmental factors, etc.) that initiated the security process, triggered rules, issued alerts, warnings (authenticated, Assessment of unauthenticated) is provided to the learning system 140. The opinion may be used to influence the 'strength value' associated with the assessment (consistent, inconsistent), or subsequent alert notification by the learning system 140, which is further discussed later. It may also include.
[26] 3 shows an exemplary flow diagram for updating a set of rules established through a learning system in accordance with the present invention. Exemplary reasoning system 150 is shown in FIG. 3 as including an external interface 310, neural network 320, and thresholder 330. External interface 310 receives item and personal identification from detectors 110 and 120 of FIG. 1, provides alerts to security personnel, converges comments based on the alerts, and the like. In the example of FIG. 3, neural network 320 is shown to achieve the 'inference' operation of inference system 150. Previously, neural network 320 includes a network of nodes that links a set of input promoters to a set of output results. Each node in the network includes a set of 'weights' applied to each input to the node, the weighted combination of input values determining the output value of the node. In this example embodiment, the learning system 140 enforces accurate security alert decisions (warnings resulting in 'unauthorized' movement decisions) and alerts that result in inaccurate security alert decisions ("authenticated" movement decisions). The feedback from the external interface 310 of the inference system 150 is processed to adjust the weight of the node to reduce the likelihood of providing. As discussed above, the opinion may include a factor that determines how strongly the particular opinion information should affect the weight of the node within the neural network 320. For example, certain expensive items may require a formal authentication process, such as an administrator signature on a form or registration in the security rules database 145. An "unauthenticated" opinion to a learning system for a person who is authorized to move an item in another way but fails to follow a formal certification process, is generally "unauthorized" about a person who is not actually authorized to move the item. It is configured to affect the node weight of the neural network 320 less than the opinion. In a similar manner, the cost of an item, or the status of an individual within an organization's rank, can be used by the learning system 140 to determine the outcome of an opinion on node weights.
[27] A typical neural network 320, or another system used to determine output based on multiple inputs, is coupled with a threshold device 330 that provides a determination as to whether the generated output allows for triggering an alert. Neural network 320 may be configured to provide a set of likelihood estimates for parameters that are assumed to be related to whether theft occurs. Threshold device 330 processes this rather independent output to determine whether to issue an alert. Common to the prior art, and as its name suggests, the threshold device 330 may include a set of thresholds for each parameter, and may trigger an alert if any parameter exceeds that threshold. have. Alternatively, threshold device 330 may form one or more mixtures of parameter values, comparing each mixture to a constant threshold. Commonly, fuzzy-logical systems are used within critical systems. As shown in FIG. 3, the example learning system 140 may provide input from the reasoning system 150 to influence thresholds, further perform accurate reasoning, and / or reduce inaccurate reasoning. Can also be used. In this way, a genetic algorithm can be used to determine effective parameters and thresholds based on an evaluation of the effectiveness of previous generations of parameters and thresholds.
[28] The overall effect of the learning system 140 is a set of input events (rule set 145) that triggers an alert by refine the rule set 145 or improve the results generated by the rule set 145. In the "+" sign will eventually be highly correlated with events that indicate potential theft, and a set of input events ("-"} in the rule set 145) will be authenticated unless the alert is triggered. It has a high correlation with. In this way, the number of alerts that need to be handled by security personnel is potentially reduced and potentially concentrated on actual security-assurance events.
[29] Similar to seasoned security staff, security systems and learning systems are configured to learn which events are "ordinary" or "usual" and thus "extra-ordinary" or "uncommon." Note that the "unusual" event is easily apparent. For example, in a home environment, a security system can be configured to define and refine rules based on consistent behavior. If someone in the family regularly takes the trombone out of their home every Thursday morning for a trombone lesson in the afternoon, the learning system can create a 'rule' that correlates to these events. On a subsequent Thursday morning, if a person leaving home without a trombone is detected, the system may issue an alert based on this 'unmatched' event. In this example, the security system uses a notification device, such as an intercom speaker at the exit, to warn the person that he has no trombone. In a similar manner, in an office environment, if a person brings an umbrella to work in the morning, the security system may remind the person to take the umbrella home in the afternoon.
[30] Various techniques may be used to achieve detection of inconsistency events. In a preferred embodiment, Bi-directional Associative Memory (BAM) is used, where a person, a person's privileges, an object, an environment (i.e., day of year, day of week, time of day, The parameters describing the temperature, etc., and the location are encoded into a vector representation suitable for input into the BAM. The BAM is then trained to recognize this pattern, preferably using a gradient search method. The selected pattern is a pattern representing a steady state; Techniques common to the prior art can be used to automate the identification of 'normally' or frequently occurring events and correlate factors associated with these events. As known in the art, BAMs are particularly well suited to determining the closest vector to be included in an input vector in a BAM. In this example, the vector in the BAM represents the normally observed state and the input vector represents the currently sensed state. If the current sensed state corresponds to a normal state, the vector closest to this current sensed state in the BAM will match the input vector. If the current detected state corresponds to an abnormal state, the closest vector in the BAM will not match the input vector. In this example, if one or two parameters in the currently sensed state do not match the encoding of a particular steady state, but if a significant number of other parameters match this particular steady state, then this steady state is identified as the nearest vector. Incorrectly matched parameters will identify abnormal events.
[31] The learning system process is represented by blocks 250 and 260 in FIG. In step 250 the opinions are converged and in step 260 the security rules are updated. 4 shows an example flow diagram corresponding to update 260 of a security rule. As shown in FIG. 4, in a preferred embodiment, different types of opinion are supported at step 415. In this example, three types of opinions are described: 'regular' opinions, 'considered' opinions, and 'returned' opinions. As will be apparent to those skilled in the art, other types of opinions, and combinations of types of opinions, may also be supported. In this example, for example, the 'periodical' opinion is the result of a random sampling test in response to a warning or in the absence of a warning. In this example embodiment, the periodic opinion only affects the threshold used to trigger the alert in step 420. On the other hand, the 'considered' opinion may be an opinion generated based on a thorough review of the transaction log, or input of opinions by senior security officials. Because the 'considered' comments are assumed to be more reliable than the 'regular' comments, the learning system uses the 'considered' comments to update the rule set at step 430. On the other hand, the overturned opinion may supercedes the existing rules at step 440 and may be provided for a typically limited duration during an emergency. Other types of comments may also be used, such as 'administrative' comments, 'administrative' comments, etc. For example, new employees may be authorized to move certain items, and senior employees may not move any items, and so forth. As discussed above, other types of opinion that are not security related may also be supported, such as when a person arrives in the surveillance area, a 'message' type that may be used to send a message to the person or items associated with the person.
[32] Note also that a paradigm of rule-based system is also provided to facilitate understanding. Other configurations and techniques are also possible. For example, reasoning system 150 may be "agent based", with each actor representing an item or an individual. Individual actors each have an initial set of rules and have the ability to learn behaviors such as regular entry and exit procedures so that they can recognize and notify of unusual behavior. The item actor has the ability to inspect in the database an individual authorized to move the item, or to initiate an account log procedure. An actor may also be designed to work with other actors. For example, one item may be an 'authentication pass' wherein the item actor is a 'authentication actor'. The authenticator acts to prevent or reduce the likelihood that a normally generated alert will not have the simultaneous provision of authentication passes.
[33] The following example illustrates a general scenario that may be supported by the system described above.
[34] The exemplary system uses the following parameters whenever an object containing one of the proximity_triggering ID tags enters or leaves the security facility: item_ID, person_ID (optional), day_of_week, time, And collect entry / exit codes.
[35] The example system also divides an event into two regions, allowed events and disallowed events. This can be accomplished by having a set of rules that distinguish between allowed and disallowed events, for example, rules that are prepared and maintained by security personnel.
[36] In order to provide the ability to build up a picture of the "normal" allowed event, even though special notifications are not allowed, special notifications may be issued when rare events occur, and the following steps are performed:
[37] 1. Define event similarity measure. For example, a template of "normal event" may be defined as any set of at least K events that share at least M characteristics. In the above-described 'thrombone' example, the event history may indicate a K event as item_ID = Trombone, Person_ID = Hugo, Date_Date = Thursday, and Type = Outing.
[38] 2. Define algorithms to define fuzzy family membership functions that capture patterns in features that do not match exactly. An example of such a fuzzy family member function would be:
[39] 2a) for items by category (ie, item_ID), their values are observed to form an item_ID set;
[40] 2b) for everyday items (ie date_date), bracketing the intervals of observed values to form a defined range;
[41] 2c) For successive items (i.e. time), it is possible to define a triangular family member function that has a peak at the mean of the observed values and is zero at slightly smaller intervals outside of the observed extreme values. have. In the example of a trombone, the distribution of time that Hugo goes out on Thursday with his trombone can be observed to have a middle of 18:30, with any observed value outside the interval between 18:17 and 18:35. Don't have
[42] 3. Define one or more less defined event similarity measures that will be used to compare new events with conventional event templates. An example may be matched on at least n-1 features, where n is the number of features that define the conventional event template described above. In the trombone example, the observed event consisting of person_ID = hugo, date_date = Thursday, type = going out, time = 18: 20 and item_ID = null matches the fuzzy one-way criterion for this less limited similarity measure However, it is different from the normal event template (there is no item_ID for the trombone).
[43] 4. Prescribe notifications to be sent in accordance with conventional event similarity measures and less defined event similarity measures. For example, if the different item is Item_ID, a warning is issued indicating that the item has been forgotten.
[44] As can be seen, definitions such as "generic" definitions and rules, i.e. "at least n-1 features" to define less defined events, and "if less limited events but not ordinary events, By providing a rule such as "Send alert of forgotten item if item_ID does not match," the system according to the present invention may provide an alert corresponding to a specific event that is not fully encoded in the rules database. In contrast, in conventional database systems, certain rules regarding each item, for example, a trombone, need to be explicitly included in the database.
[45] The foregoing description merely illustrates the principles of the invention. Thus, it will be appreciated that those skilled in the art may devise various arrangements which, although not explicitly described or shown herein, embody the principles of the invention and are therefore within the spirit and scope of the invention. For example, the advantages provided by the learning system to change security rules based on comments from security events can be achieved regardless of the means used to identify items and / or individuals. That is, conventional card readers, UPC code readers, biographical scanners, pattern recognition systems, image processing systems, and the like may form detectors 110 and 120 used to identify items or individuals. In a similar manner, the advantages provided by the use of a remote transponder can be achieved regardless of the means used to maintain or update the rules in force. That is, for example, a conventional database management system may be used by the inference system 150 for associating an item with an individual authorized to move the item, or a system based on conventional rules may be a learning system 140. Can be used without using In a similar manner, although a security system may be present herein as a system that restricts unauthorized movement of items from a security facility, the system may also be used to restrict the entry of unauthorized items into the security facility. For example, if a transponder is instructed to be installed on all fire extinguishers, the system can be used to prohibit the transport of fire extinguishers to secure areas except authorized personnel. Such system configurations and other system configuration and optimization features and other configurations and features will be apparent to those skilled in the art in view of this disclosure and are included within the scope of the following claims.
[46] As described above, the present invention relates to the field of security systems, and in particular, it is used in security systems and the like for adaptively generating and changing security rules and parameters based on conventional events.

权利要求:
Claims (13)
[1" claim-type="Currently amended] An item detector 120 configured to detect the identified item 102, and
An individual detector 110 configured to detect the identified person 101,
Generate an alert in accordance with the identified item 102, the identified person 101, and a set of security rules 145,
To collect feedback in response to the above warning;
Configured, a reasoning system 150,
Learning system 140 configured to modify the set of security rules 145 in accordance with the comment.
Including, security system.
[2" claim-type="Currently amended] The method of claim 1 wherein each of the identified items 102 and the identified persons 101 has an associated transponder with a unique unit identification,
The item detector (120) and the personal detector (110) each comprise a single detector unit configured to detect the unit identification from an associated transponder.
[3" claim-type="Currently amended] The method of claim 1, wherein at least one of the item detector 120 and the personal detector 110,
With card reader,
A biometric device,
An image processing device,
A pattern recognition device,
Transponder detector
A security system comprising at least one of.
[4" claim-type="Currently amended] The system of claim 1, wherein the learning system 140 comprises a neural network, an expert system, an agent system, an associative memory, a genetic algorithm, and fuzzy logic. At least one of a system and a rule-based system.
[5" claim-type="Currently amended] The method of claim 1,
The learning system 140 is further configured to change the set of rules in accordance with at least one other parameter associated with the alert,
The at least one other parameter is,
A time of day,
A day of week,
Temperature,
The direction of movement of at least one of the identified item 102 and the identified person 101,
The presence of other identified items,
The presence of other identified persons,
At least one of the security
Including, security system.
[6" claim-type="Currently amended] The method of claim 1,
The comments include class-type,
The learning system 140 is further configured to change the set of rules according to the class-type of the opinion,
The class-type includes at least one of routine, considered, temporary, absolute, and override.
[7" claim-type="Currently amended] Detecting 210 the presence of the identified item 102;
Detecting (220) the presence of the identified person (101),
Generating (240) an alert in accordance with the identified item (102), the identified person (101), and a set of security rules (145);
Collecting 250 an opinion associated with the alert;
Automatically changing (260) the set of security rules (145) based on the comment.
Including, security method.
[8" claim-type="Currently amended] The method of claim 7, wherein
Each of the identified item 102 and the identified person 101 has an associated unique identifier,
Detecting the presence of at least one of the identified item 102 and the identified person 101,
Receiving the unique identifier from a transponder associated with at least one of the identified item 102 and the identified person 101;
Reading the unique identifier from a card associated with at least one of the identified item 102 and the identified person 101;
Processing an image corresponding to at least one of the identified item 102 and the identified person 101,
At least one of reading a characteristic embedded in at least one of the identified item (102) and the identified person (101) to determine the associated unique identifier.
[9" claim-type="Currently amended] The method of claim 7, wherein
The step 260 of automatically changing the set of security rules 145 comprises at least one of a neural network, an expert system, an actor system, an associative memory, a genetic algorithm, a fuzzy logic system, and a rule based system. Security method, including the use of one.
[10" claim-type="Currently amended] The method of claim 7, wherein
Automatically changing (260) the set of security rules (145),
Vision,
Date,
Temperature,
The direction of movement of at least one of the identified item 102 and the identified person 101,
The presence of other identified items,
The presence of other identified persons,
And further based on at least one of the security states.
[11" claim-type="Currently amended] The method of claim 7, wherein
The step 260 of automatically changing the set of security rules 145 is further based on the class-type associated with the comment,
The class-type includes at least one of periodic, contemplated, temporary, absolute, and reversal.
[12" claim-type="Currently amended] As security system 100,
Emit one or more trigger signals,
Receive two or more responses from one or more of the trigger signals from two or more transponders remote from the detector, one of the two or more responses corresponding to an individual's identification and the other of the two or more responses To receive the two or more responses corresponding to the identification of the item
Detectors 110 and 120 are configured,
Reasoning system 150, configured to provide a security event in accordance with the identification of the individual and the identification of the item, operatively coupled to the detectors 110, 120;
Provide notification of the security event to security personnel,
Receive feedback from the security personnel based on the notification;
A security interface 130 configured and operatively coupled to the inference system 150,
And based on the comments received from the security personnel based on the notification, influence the determination of the reasoning system 150 for subsequent security events, and the reasoning system 150 and the security interface 130 Learning system 140, operatively coupled to
Including, security system.
[13" claim-type="Currently amended] The method of claim 12,
Additionally includes a set of security rules 145,
The learning system (140) is configured to influence the determination of the reasoning system (150) for the subsequent security event by changing the set of security rules (145).

类似技术:

---

公开号 | 公开日 | 专利标题

---

EP3149986B1|2019-12-11|Electromagnetic persona generation based on radio frequency fingerprints

---

US10026302B1|2018-07-17|Mobile alarm device

---

US9070233B2|2015-06-30|Automated banking machine system and monitoring

---

US10671821B2|2020-06-02|Systems and methods for RFID-based access management of electronic devices

---

Govindarajan et al.2011|Intrusion detection using neural based hybrid classification methods

---

US6331817B1|2001-12-18|Object tracking apparatus and method

---

US20140118140A1|2014-05-01|Methods and systems for requesting the aid of security volunteers using a security network

---

US6774782B2|2004-08-10|Radio frequency personnel alerting security system and method

---

US8681984B2|2014-03-25|Context-aware based RFID privacy control system and personal privacy protection method using the same

---

EP1991850B1|2014-09-10|Rfid/biometric area protection

---

US20120270560A1|2012-10-25|Methods to authenticate access and alarm as to proximity to location

---

JP2017511544A|2017-04-20|Person authentication and tracking system

---

WO2016178082A1|2016-11-10|Authentication channel flow through wearable

---

AU2012321152B2|2017-11-23|Sanitization protocol monitoring/compliance systems, apparatuses, methods, and software

---

JP4142062B2|2008-08-27|Monitoring system and terminal device

---

ES2431667T3|2013-11-27|Device and method to detect suspicious activity, program and recording medium

---

US20130278390A1|2013-10-24|Systems and methods for object localization and path identification based on rfid sensing

---

US8700769B2|2014-04-15|System and method for providing configurable security monitoring utilizing an integrated information system

---

JP4559819B2|2010-10-13|Suspicious person detection system and suspicious person detection program

---

Mirowski et al.2007|Deckard: A system to detect change of RFID tag ownership

---

US10503921B2|2019-12-10|Emergency notification system and methods

---

EP3243191B1|2019-05-22|Anti-theft system used for customer service

---

KR100194987B1|1999-06-15|Entry and exit monitoring system

---

US20120237908A1|2012-09-20|Systems and methods for monitoring and managing use of mobile electronic devices

---

US6437696B1|2002-08-20|Prisoner tracking and warning system and corresponding methods

同族专利:

---

公开号 | 公开日

---

JP2003536184A|2003-12-02|

---

US6300872B1|2001-10-09|

---

EP1297508A2|2003-04-02|

---

WO2001099075A2|2001-12-27|

---

US6492905B2|2002-12-10|

---

WO2001099075A3|2002-04-18|

---

US20010052851A1|2001-12-20|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

法律状态:
2000-06-20|Priority to US09/597,197

---

2000-06-20|Priority to US09/597,197

---

2001-06-15|Application filed by 요트.게.아. 롤페즈, 코닌클리케 필립스 일렉트로닉스 엔.브이.

---

2001-06-15|Priority to PCT/EP2001/006888

---

2002-04-18|Publication of KR20020029382A

优先权:

---

申请号 | 申请日 | 专利标题

---

US09/597,197|US6300872B1|2000-06-20|2000-06-20|Object proximity/security adaptive event detection|

---

US09/597,197|2000-06-20|

---

PCT/EP2001/006888|WO2001099075A2|2000-06-20|2001-06-15|Object proximity/security adaptive event detection|