Building access control system with spatial modeling

1. A method in an access control system, the method comprising:

maintaining a database of access control event data generated by a plurality of access control devices installed in a building;

iterating through the access control event data in order to generate a connectivity model for the building by, for each iteration:

identifying an interaction between a user and an access control device for a first door in the building that occurs at a first time;

identifying an interaction between the user and an access control device for a second door in the building that occurs at a second time;

determining if a difference between the first time and the second time is less than a threshold period of time;

determining if the connectivity model includes a connection between the first door and the second door responsive to a determination that the difference is less than the threshold;

creating the connection between the first door and the second door responsive to a determination that the connectivity model does not include the connection; and

updating a weight associated with the connection responsive to a determination that the connectivity model does include the connection.

2. The method of claim 1 , further comprising removing or disregarding the connection from the connectivity model based on the weight.

3. The method of claim 2 , wherein removing or disregarding the connection from the connectivity model based on the weight comprises removing or disregarding the connection from the connectivity model in response to the weight being less than a threshold.

4. The method of claim 2 , wherein the connection comprises a first connection and the weight comprises a first weight, and wherein removing or disregarding the first connection from the connectivity model based on the first weight comprises removing or disregarding the first connection from the connectivity model based on a comparison of the first weight to a second weight associated with a second connection.

5. The method of claim 4 , further comprising calculating a weight associated with each connection in the connectivity model and removing or disregarding the connection in response to the weight falling below a threshold percentage within a distribution of the weights across the connections in the connectivity model.

6. The method of claim 4 , further comprising:

calculating a weight associated with each connection in the connectivity model;

applying a model to the connections to separate the connections into a plurality of clusters of connections based on similarities of weights;

determining a first cluster of the plurality of clusters having smallest weights; and

removing or disregarding the connections belonging to the first cluster.

7. The method of claim 6 , wherein the model comprises a multi-state Poisson mixture model, and wherein determining the first cluster comprises identifying the first cluster as a cluster of the plurality of clusters having a smallest Poisson rate parameter from among a plurality of Poisson rate parameters of the plurality of clusters.

8. The method of claim 1 , further comprising ignoring the interaction between the user and the first door and the interaction between the user and the second door responsive to a determination that the difference is greater than the threshold.

9. The method of claim 1 , further comprising identifying an access control zone within the building using the connectivity model.

10. The method of claim 9 , further comprising generating an alarm that indicates an intrusion associated with the building responsive to multiple door forced open events occurring within the access control zone within a time period.

11. The method of claim 1 , further comprising estimating a probability of a first user coming into contact with a second user or a restricted location using the connectivity model based on a last access control device with which the first user interacted.

12. An access control system comprising:

one or more processors; and

one or more computer-readable storage media having instructions stored thereon that, when executed by the one or more processors, cause the one or more processors to implement operations comprising:

maintaining a database of access control event data generated by a plurality of access control devices installed in a building;

iterating through the access control event data in order to generate a connectivity model for the building by, for each iteration:

identifying an interaction between a user and an access control device for a first door in the building that occurs at a first time;

identifying an interaction between the user and an access control device for a second door in the building that occurs at a second time;

determining if a difference between the first time and the second time is less than a threshold period of time;

determining if the connectivity model includes a connection between the first door and the second door responsive to a determination that the difference is less than the threshold;

creating the connection between the first door and the second door responsive to a determination that the connectivity model does not include the connection; and

updating a weight associated with the connection responsive to a determination that the connectivity model does include the connection.

13. The system of claim 12 , the operations further comprising removing or disregarding the connection from the connectivity model based on the weight.

14. The system of claim 13 , wherein removing or disregarding the connection from the connectivity model based on the weight comprises removing or disregarding the connection from the connectivity model in response to the weight being less than a threshold.

15. The system of claim 13 , wherein the connection comprises a first connection and the weight comprises a first weight, and wherein removing or disregarding the first connection from the connectivity model based on the first weight comprises removing or disregarding the first connection from the connectivity model based on a comparison of the first weight to a second weight associated with a second connection.

16. The system of claim 15 , the operations further comprising calculating a weight associated with each connection in the connectivity model and removing or disregarding the connection in response to the weight falling below a threshold percentage within a distribution of the weights across the connections in the connectivity model.

17. The system of claim 15 , the operations further comprising:

calculating a weight associated with each connection in the connectivity model;

applying a model to the connections to separate the connections into a plurality of clusters of connections based on similarities of weights;

determining a first cluster of the plurality of clusters having smallest weights; and

removing or disregarding the connections belonging to the first cluster.

18. The system of claim 17 , wherein the model comprises a multi-state Poisson mixture model, and wherein determining the first cluster comprises identifying the first cluster as a cluster of the plurality of clusters having a smallest Poisson rate parameter from among a plurality of Poisson rate parameters of the plurality of clusters.

19. The system of claim 12 , the operations further comprising ignoring the interaction between the user and the first door and the interaction between the user and the second door responsive to a determination that the difference is greater than the threshold.

20. The system of claim 12 , the operations further comprising identifying an access control zone within the building using the connectivity model.

21. The system of claim 20 , the operations further comprising generating an alarm that indicates an intrusion associated with the building responsive to multiple door forced open events occurring within the access control zone within a time period.

22. The system of claim 12 , the operations further comprising estimating a probability of a first user coming into contact with a second user or a restricted location using the connectivity model based on a last access control device with which the first user interacted.