Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, are described for performing privacy-aware surveillance with a security camera drone. The drone obtains image data using an imaging device and processes the image data to detect an object external to the drone. In response to processing the image data, the drone determines multiple viewing angles indicated in the image data with respect to the object. Based on the image data and the multiple viewing angles, a first section of the area for surveillance and a second, different section of the area to be excluded from surveillance are identified. The drone determines an adjustment to at least one of the multiple viewing angles to cause the second section to be excluded from surveillance. Based on the adjustment, the drone controls the imaging device to exclude the second section from surveillance imagery obtained by the drone.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for property communication and access control. In some implementations, a corresponding method includes receiving a data sequence input from a property related to an event at the property; comparing the data sequence input to one or more sequence templates to determine a matching sequence template; based on the matching sequence template, determining one or more relevant actions to be performed; performing the one or more relevant actions using the data sequence input; and providing feedback related to the event or the one or more relevant actions.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a storage device, for a monitoring system that is configured to monitor a property. The monitoring system may include a sensor that is configured to generate sensor data that reflects an attribute of the property, a microphone that is connected to a pipe and that is configured to detect audio data associated with the pipe, and a monitor control unit. The monitor control unit is configured to receive the sensor data and the audio data, provide the sensor data and the audio data as inputs to a model that is trained to determine that a device connected to the pipe is being used, receive, from the model, data indicating that the device connected to the pipe is being used, and based on determining that the device connected to the pipe is being used, perform a monitoring system action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a non-transitory computer storage medium, for receiving user-defined routes, each user-defined route including a tripwire of multiple tripwires and a direction, where each user-defined route includes an action of multiple actions executed by a sub-system of a home monitoring system, detecting a first signal at a first tripwire including a first direction, determining a set of user-defined routes including the first tripwire and the first direction, detecting a second signal at a second tripwire including a second direction, determining a particular user-defined route of the set of user-defined routes including the first tripwire and the first direction and the second tripwire and the second direction, determining an action executed by a sub-system of the home monitoring system for the particular user-defined route traversed by a user, and triggering execution of the action based on the second signal at the second tripwire.

Abstract: Methods, systems, apparatus, and computer programs, for tracking objects are disclosed. In one aspect, a method is disclosed that includes actions of obtaining an image, determining that a user of a first monitoring system has opted-in for object tracking by a second monitoring system that is remote from the first monitoring system, and based on a determination that the user of the first monitoring system has opted-in for object tracking: determining whether the obtained image satisfies a predetermined level of similarity to a stored tracking object image model stored on a first device of the first monitoring system, and based on a determination that the obtained image satisfies a predetermined level of similarity to the stored tracking object image model, generating a tracking update notification, and transmitting the tracking update notification to the second monitoring system that is remote from the first monitoring system.

Abstract: Systems and techniques are provided for identifying patterns activity at a facility that seem to precede alarm events. In some implementations, a monitoring system is configured to monitor a property and includes a sensor configured to generate sensor data. A monitor control unit is configured to receive the sensor data and based on the sensor data, determine an alarm status of the monitoring system. A monitoring application server is configured to receive the sensor data and data indicating the alarm status of the monitoring system. The server applies the sensor data to a model that determines an estimated alarm status of the monitoring system and determines the estimated alarm status of the monitoring system is different from the alarm status. Based on determining that the estimated alarm status of the monitoring system is different from the alarm status of the monitoring system, the server overrides the alarm status.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, are described for implementing a smart firearm safety device. The safety device attaches to a firearm having a trigger and a slot for receiving a magazine. The safety device includes a locking mechanism that attaches to the trigger to preclude depressing a trigger of the firearm and a sensor that determines an orientation of the firearm or a relative motion of the firearm to indicate detected movement of the firearm. The safety device also includes a radio device that receives parameter signals from the sensor indicating movement of the firearm. The radio device communicates with a component of a property monitoring system to receive a command to engage the locking mechanism to preclude depressing the trigger of the firearm based on parameter signals indicating a particular type of detected movement of the firearm.

Abstract: A monitoring system that is configured to monitor a property is disclosed. The monitoring system includes a sensor that is configured to generate sensor data that reflects an attribute of a property. The monitoring system further includes a drone that generates image data, location data, and orientation data. The monitoring system further includes a monitor control unit. The monitor control unit is configured to receive the sensor data, the location data, and the orientation data. The monitor control unit is configured to determine that an event has occurred at the property and a location of the event within the property. The monitor control unit is configured to generate a graphical overlay based on the event, the location data, and the orientation data. The monitor control unit is configured to generate a graphical interface. The monitor control unit is configured to output the graphical interface.

Abstract: Techniques are described for distributing, to a distributed network of central stations, alarm events detected in monitoring system data collected by sensors included in monitoring systems located at monitored properties. A system receives monitoring system data collected by sensors included in monitoring systems located at monitored properties, tracks alarm events detected within the monitoring system data, and generates, for central station servers in a distributed network of central stations, load profiles that reflect a volume of alarm events being handled at each of the central station servers at a particular period of time. The system determines capacities to handle additional alarm events for the central station servers, determines relative priorities for the central station remote servers based on the determined capacities, and directs subsequent alarm events to the central station servers based on the relative priorities.

Abstract: Techniques and systems are described for creating and operating a vibration triangulation network. The network can include vibration sensor devices placed in a building, such as a store or a house, which are able to pinpoint the location of a target, e.g., person, animal, etc., through triangulation. A described technique includes receiving vibration notifications from vibration sensor devices, determining a location of a target within the building based on at least a portion of the vibration notifications and a vibration sensor map associated with the building, and triggering an alarm event based on a detection of the target.

Abstract: Methods, systems, and apparatus for extended training data sets for neural network-based IT solutions. In one aspect, a method includes collecting, support data generated during multiple support events including a corpus of articles referenced during the support event. Within the corpus of articles, restricted and unrestricted articles referenced in the support data and restricted and unrestricted articles not referenced in the support data are identified. Embedded vectors are generated for each article that is referenced in the support data from the article and a subset of the support data that references the article, and for each article that is not referenced by the support data, an embedded vector from only the article. A dimensionality of the embedded vectors is reduced and a neural network is trained using the embedded vectors to select a particular article of the corpus of articles responsive to a new support event.

Abstract: Methods, systems, and apparatus for video analytics evaluation are disclosed. A method includes: identifying a video to display a result of an evaluation of video analysis; identifying a particular time in the video when a video analysis determination does not match a ground truth determination for the video; displaying an image from the particular time in the video; and displaying an indication that the video analysis determination does not match the ground truth determination for the video. Displaying the image from the particular time in the video can include generating a graphical user interface for presentation on a display of a computing device. The indication that the video analysis determination does not match the ground truth determination for the video can include a user-selectable icon. In response to a user selecting the user-selectable icon, the method can include displaying video analysis results for the particular time in the video.

Abstract: A monitoring system that is configured to monitor a property is disclosed. The monitoring system includes a sensor that is configured to generate sensor data that reflects an attribute of the property; a solar panel that is configured to generate and output power; and a monitor control unit. The monitor control unit is configured to: monitor the power outputted by the solar panel; determine that the power outputted by the solar panel has deviated from an expected power range; based on determining that the power outputted by the solar panel has deviated from the expected power range, access the sensor data; based on the power outputted by the solar panel and the sensor data, determine a likely cause of the deviation from the expected power range; and determine an action to perform to remediate the likely cause of the deviation from the expected power range.

Abstract: Systems and techniques are described for an integrated security environment to monitor activities at a commercial facility and a residential facility. In some implementations, a monitoring system monitors a property that includes one or more sensors located at the property and generate sensor data. A monitor control unit receives and analyzes the sensor data. Based on analyzing the sensor data, the monitor control unit determines that an event has likely occurred at the property and generates data indicating that the event has likely occurred at the property. A monitoring server receives the data indicating the event has likely occurred at the property. Based on the data that the event has likely occurred at the property, transmitting, to an additional monitoring system that is configured to monitor an additional property, instructions for the additional monitoring system to perform an action.

Abstract: Mesh network resiliency technology, in which a first routing configuration for nodes of a mesh network is determined, the first routing configuration being appropriate when a first power source of the mesh network is available. Routing data that indicates routing responsibilities within the first routing configuration is provided to first nodes of the mesh network. An interruption of the first power source for the mesh network is detected by nodes of the mesh network. In response to detecting the interruption, a second routing configuration for nodes of the mesh network is determined, the second routing configuration being appropriate when the first power source of the mesh network is unavailable. Routing data that indicates routing responsibilities within the second configuration is provided to second nodes of the mesh network, each of the second nodes including a second power source that is different than the first power source.

Abstract: Techniques are described for using information from a knob sensor of an appliance in a property monitoring system. In some implementations, a monitoring system may be configured to receive sensor information from a knob, receive monitoring device information from a monitoring device and, based on the sensor information from the knob, identify a knob event. The monitoring system may integrate the knob event with the monitoring device information and analyze the integrated knob event and monitoring device information against one or more rules related to usage of the appliance in view of the state of the property. Based on analyzing the integrated knob event and monitoring device information, the monitoring system may determine whether to perform an action, such as notifying a user or activating an appliance.

Abstract: Techniques are described for detecting and handling unauthenticated commands in a property monitoring system. In some implementations, a monitoring system may include sensors located throughout a property, a monitoring control unit, and an input device. The monitoring control unit may be configured to receive data collected by the sensors, as well as an input command detected by the input device. For an input command that does not include authentication information, the monitoring control unit may generate property state information based on the sensor data, then analyze the property state data and the input command against one or more rules that relate to authorization of unauthenticated commands. Based on the analysis, the monitoring control unit may determine whether to perform the action corresponding to the input command or whether to perform another action, for example, generating and providing a notification or authorization request to a user.

Abstract: Methods, systems, and apparatus for monitoring humidity at a property are disclosed. A monitoring system performs operations including: obtaining, from multiple humidity sensors within the property, humidity data; obtaining occupancy data that indicates whether people are within the property; determining, based on the humidity data and the occupancy data, that humidity in a first area of the property changed more relative to a change in humidity in a second area of the property during a time period when no person was present in the first area and no person was present in the second area; and in response to determining that humidity in the first area changed more relative to a change in humidity in a second area of the property during the time period, providing an indication of a water leak in the first area of the property to a user.

Abstract: Techniques are described for using sensor-based objection recognition and property condition monitoring techniques to automate and improve methods of navigating a property during a property tour. In some implementations, sensor data is received from one or more sensors that are located within a property. Property data that identifies characteristics of the property is received. A model of the property is generated. Authentication data that identifies users who are authorized to access the property is received. Data identifying a visiting user is received. The data identifying the visiting user is compared to the authentication data. The visiting user is determined to be authorized to access the property. Preference data that indicates preferences of the visiting user is accessed. A path to guide the user around the property is generated based on the model of the property and the preferences of the visiting user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for integrating one or more light switch panels into a monitoring system. The methods, systems, and apparatus include actions of determining to communicate a non-traditional file to a z-wave capable device, communicate a first data packet to the z-wave capable device, based on communicating the first data packet to the z-wave capable device, alter a schedule command class payload of a second data packet to include a first portion of the non-traditional file, alter a schedule command class payload of a third data packet to include a second portion of the non-traditional file, and alter a schedule command class payload of a fourth data packet to include a third portion of the non-traditional file, and communicate, to the z-wave capable device, the altered second data packet, the altered third data packet, and the altered fourth data packet.