# US 2020/0394892 A1 · Detecting Abnormal Events for a Monitoring System

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0394892 A1
- Document ID: US20200394892A1
- Status: Published application
- Kind: A1
- Application number: 16/439,094
- Filed: 2019-06-12
- Published: 2020-12-17
- Earliest claimed: 2019-06-12
- Assignees: Comcast Cable Communications, LLC
- Applicants: Comcast Cable Communications, LLC
- Inventors: Zhen Zhao
- Web page: https://patentgrep.com/patent/US20200394892A1

## CPC

- G08B 21/24: Reminder alarms, e.g. anti-loss alarms
- G06F 16/219: {Managing data history or versioning (querying versioned data G06F16/2474; querying temporal data G06F16/2477)}
- G08B 13/08: by opening, e.g. of door, of window, of drawer, of shutter, of curtain, of blind
- G08B 25/008: {Alarm setting and unsetting, i.e. arming or disarming of the security system}
- G08B 13/00: Burglar, theft or intruder alarms
- G08B 25/14: Central alarm receiver or annunciator arrangements
- G08B: SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Systems, apparatuses, and methods are described for detecting a state transition for sensors of a monitoring system, and determining whether the state transition is abnormal such that further action is warranted. The determination may be based on a history of state transitions in the monitoring system, and may be based on whether the same state transition has previously occurred at a corresponding time in the history, whether the new state has previously occurred following the corresponding time in the history, and other factors.

## Claims

### Claim 1

A method comprising:
maintaining state history information associated with sensors that are associated with a premises;
detecting a current state transition in which at least one of the sensors changes from a first state to a second state;
determining, from the state history information, one or more prior state transitions in which the at least one of the sensors changed from the first state to the second state;
determining, from the state history information, time windows corresponding to a time of the current state transition; and
sending, based on the one or more prior state transitions and the time windows, an anomalous activity alert.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
determining, based on the time of the current state transition, prior time intervals from the state history information, wherein the prior time intervals are periodically repeated within a time range of the state history information; and wherein the sending the anomalous activity alert is further based on determining, based on the current state transition, whether the one or more prior state transitions occurred during the prior time intervals.

### Claim 3 (depends on 1)

The method of claim 1, wherein the time windows comprise a first time window centered at a time that is aligned with the time of the current state transition and second time windows located before and after the first time window, wherein the method further comprises adjusting a security score based on a determination of whether the one or more prior state transitions occurs in the first time window or the second time windows.

### Claim 4 (depends on 1)

The method of claim 1, wherein the determining the time windows corresponding to the time of the current state transition further comprises determining that the time windows comprise one or more prior state transitions that match the current state transition; and
wherein the method further comprises applying one or more weighted scores to the one or more prior state transitions.

### Claim 5 (depends on 4)

The method of claim 4, wherein the one or more weighted scores decrease for time windows that are farther, in time, from a time that is aligned with the time of the current state transition.

### Claim 6 (depends on 1)

The method of claim 1, wherein the sending the anomalous activity alert is further based on:
a state transition score that is based on prior state transitions in which the at least one of the sensors changed from the first state to the second state; and
a new state score that is based on times, in the state history information, in which the at least one of the sensors was in the second state.

### Claim 7 (depends on 6)

The method of claim 6, wherein the new state score is further based on:
determining one or more sections of one or more time windows in which the at least one of the sensors was in the second state; and
adjusting, based on recency of the second state, the new state score.

### Claim 8 (depends on 1)

The method of claim 1, wherein the state history information indicates a plurality of previous states and corresponding recorded times, wherein the previous states and the corresponding recorded times are associated with the sensors.

### Claim 9

A method comprising:
maintaining state history information based on inputs from one or more sensors associated with a premises;
determining, from the state history information, a current state transition in which the one or more sensors change from a first state to a second state;
determining, based on a time of the current state transition, time windows within the state history information;
determining, from the state history information, one or more sections of the time windows in which the one or more sensors were in the second state; and
sending, based on the current state transition and the one or more sections of the time windows, an anomalous activity alert.

### Claim 10 (depends on 9)

The method of claim 9, wherein the sending the anomalous activity alert comprises determining that a security system comprising the one or more sensors is unarmed.

### Claim 11 (depends on 9)

The method of claim 9, further comprising:
determining, based on a time of the current state transition, prior time intervals from the state history information, wherein the prior time intervals are regularly repeated for a duration of the state history information, and wherein the sending the anomalous activity alert is further based on determining whether the second state occurred during the prior time intervals.

### Claim 12 (depends on 11)

The method of claim 11, wherein the time windows are determined based on the prior time intervals determined from the state history information.

### Claim 13 (depends on 9)

The method of claim 9, wherein the time windows comprise a first time window centered at a time that is aligned with the time of the current state transition and second time windows located before and after the first time window, and wherein the method further comprises adjusting a security score based on a determination of whether the second state occurs in the first time window or the second time windows.

### Claim 14 (depends on 9)

The method of claim 9, wherein the determining the time windows within the state history information further comprises determining that the time windows comprise one or more prior state transitions that match the current state transition, and wherein the method further comprises applying one or more weighted scores to the one or more prior state transitions.

### Claim 15 (depends on 9)

The method of claim 9, wherein the sending the anomalous activity alert is further based on a security score that is based on:
one or more state transitions in the state history information that match with the current state transition; and
the one or more sections of the time windows in which the one or more sensors were in the second state.

### Claim 16 (depends on 9)

The method of claim 9, further comprising:
determining, based on the current state transition and the one or more sections of the time windows, a security score;
determining, based on a user feedback, that the current state transition has been confirmed abnormal; and
adjusting, based on the determining that the current state transition has been confirmed abnormal, the security score.

### Claim 17

A method comprising:
generating state history information indicating a state transition in which sensors associated with a premises change from a first state to a second state;
determining one or more weighted scores associated with time windows within the state history information, wherein the time windows surround a time corresponding to a time of the state transition, and wherein the one or more weighted scores are based on:
a determination of whether the state transition occurred in any of the time windows; and
a determination of a section, of at least one of the time windows, that indicates a proportion of time spent in the second state; and
determining, based on the one or more weighted scores, whether to send an anomalous activity alert.

### Claim 18 (depends on 17)

The method of claim 17, further comprises:
filtering, based on predetermined criteria, inputs from the sensors; and
encoding the inputs into a linked list.

### Claim 19 (depends on 17)

The method of claim 17, further comprising:
sending, based on the one or more weighted scores, the anomalous activity alert to authorities.

### Claim 20 (depends on 17)

The method of claim 17, wherein the time windows comprise:
a first time window aligned with the time corresponding to the time of the state transition; and
one or more additional time windows located before and after the first time window.

### Claim 21 (depends on 17)

The method of claim 17, further comprising:
adjusting, based on determining one or more sections of the time windows, the one or more weighted scores; and
sending, based on the one or more weighted scores, an anomalous activity alert.

### Claim 22 (depends on 17)

The method of claim 17, wherein the one or more weighted scores are further based on:
one or more state transitions in the state history information that match with the state transition; and
one or more sections of the time windows in which the sensors were in the second state.

### Claim 23 (depends on 22)

The method of claim 22, further comprising:
determining, based on a user feedback, that the state transition is a confirmed abnormal transition; and
adjusting, based on the determining that the state transition is the confirmed abnormal transition, the one or more weighted scores.
