Editor’s Note: This article was first published on our sister site CE Pro.
LAS VEGAS – At CES 2017, Alarm.com has announced plans to develop autonomous, video-enabled drone applications to extend the capabilities of smart home and business security systems.
The drones will use Alarm.com’s new multi-sensor awareness and property intelligence, dubbed Insights Engine, together with the Qualcomm Snapdragon Flight drone platform, to investigate unexpected activity.
The drones will be designed to autonomously navigate through the property and provide a high-resolution video feed to the property owner.
The company’s new Insights Engine learning capability, which is also being announced at CES, applies machine-learning algorithms to the growing data set generated by devices and sensors in a connected property. By recognizing complex activity patterns and detecting anomalies, the system can proactively respond to events on behalf of the user.
Alarm.com intends to leverage these insights to intelligently deploy video-enabled drones to the location where unexpected activity was detected or when an alarm was triggered. Strict, opt-in privacy controls will enable the property owner to share the video feed with a central monitoring station and emergency responders.
“This is a very interesting application for drones, which at their core are essentially flying cameras, and by using other advanced capabilities of our Snapdragon processors, Alarm.com is designing a whole new way to provide security to properties,” says Hugo Swart, senior director, product management, Qualcomm Technologies, Inc. “The Qualcomm Snapdragon Flight platform pushes the boundaries of the drone industry and has led to many new form factors and use cases, and Alarm.com, with its security expertise, is taking commercial and residential security systems to the next level by integrating intelligent and cutting edge drones to its solutions.”
“Alarm.com has been a pioneer in the smart home and business security space,” says Daniel Kerzner, Alarm.com’s chief product officer. “By analyzing data across the devices on our platform, we have created unique capabilities that make properties safer, smarter and more efficient. We’re excited to use Qualcomm Snapdragon Flight to develop similar applications for drones. Our intelligent drone deployment system will be designed to provide targeted video in response to specific events around a property – augmenting fixed location cameras and enhancing the active security perimeter for homes and businesses.”
What Is ‘Insights Engine’?
Insights Engine is the company’s new proprietary machine-learning capability that safeguards homes and businesses by identifying patterns and insights in the growing set of data generated by devices and sensors in a connected property.
By learning the unique activity patterns of any home or business, the Insights Engine can respond to unusual activity on behalf of the homeowner by taking action through Alarm.com’s broad ecosystem of connected devices.
“Alarm.com has always aggregated sensor data to create an intelligent, proactive and contextually aware smart home experience,” says Kerzner. “The Insights Engine is an example of how we are extending that approach with machine learning to define the next generation of smart home solutions.”
Unexpected Activity Notifications: The first application of the Insights Engine is a new category of intelligent awareness notifications that alert families to potential safety or security issues around their home. If the Insights Engine determines that the homeowner should be alerted, a push notification is sent, all without the homeowner having to create custom rules or notifications. For example:
Early playtime: The front door to a home is frequently opened and closed early on weekday mornings, as the family leaves for work and school. On weekends, however, it’s typically not opened until later in the day. When the door opens early on a Saturday morning, an unexpected activity notification prompts the homeowner that the kids are exploring the front yard.
Date Night: A home’s front door is typically locked by 7:30 p.m. on weekdays and 9:00 p.m. on weekends. While out at dinner, the homeowners receive a notification that the door is still unlocked at 9:15 p.m. on Saturday, and realize that the babysitter has forgotten to lock it. Acting on this information, they lock the door remotely using their Alarm.com app.
Adaptive Learning: The Alarm.com Insights Engine continually adapts based on a range of factors including user input, changes in activity patterns and external factors.
Actionable notifications allow users to “train” their smart homes to learn their preferences faster and to quickly adjust to new schedules. Notifications come with simple feedback options including Yes or No on the usefulness of each notification, enabling the Insights Engine to adapt quickly based on minimal user input. A third feedback option lets users ‘pause’ notifications for 24 hours to incorporate a sudden but temporary schedule change such as a sick day.
Time-intelligent notifications adjust in sensitivity depending on when an event takes place. An open garage door during daylight hours, for example, may not trigger an immediate notification. However, after sunset the system becomes more sensitive to security risks and is more likely to immediately alert the homeowner.