Human-in-the-Loop Robotics Demonstrates the Power of Partnership

No job is perfect. Hopefully, you love certain aspects of your job, but there are probably parts of it you’d happily ditch if you had a choice. If you’re like most people, those parts include the monotonous, routine tasks that are time-consuming to handle but must be performed with accuracy and attention to detail. What if you could offload those tedious responsibilities to someone – or something – you could trust to do the job well, so you could focus on more interesting responsibilities that require situational analysis, decision making, and human insight?

For workers in the security field, robots are now making this a reality.

Thanks to technological advancements in AI, robots now have the capability to safely navigate spaces shared by humans, such as offices and warehouses. This makes them an ideal mechanism for helping out with security patrols – specifically the parts that humans don’t usually enjoy and sometimes don’t do so well. Robots are perfect for handling the monotonous parts of the job; they perform vigilant, repetitive patrolling of defined areas and identify any anomalies that warrant further investigation.

What happens when the robots find something? That’s where the value of a human/robot partnership really shines. Robots may excel at finding abnormalities, but humans are best at assessing the context of the situation and making a decision about how to respond. An exciting trend for security robotics, for which Cobalt is leading the charge, is the integration of human-in-the-loop control rather than 100% autonomy. This innovative approach leverages what machines are best at (reliability, handling long monotonous hours and cost savings) with what humans are best at (handling outlier situations, interacting with other people and adding a friendly touch to the interaction), and combines them into a single solution. The result is a security robot that provides superior results while, at the same time, is less expensive to develop and simpler to deploy.

This approach is analogous to how we’ve seen AI impact other, more established product categories within the physical security market, where AI works in tandem with human operators to deliver the best outcome. The adoption of video analytics within the video surveillance market is a good example of this, where technology serves as a force multiplier rather than a replacement for human ingenuity. For example, when surveillance analytics engines identify a “package left behind” on a subway platform or a vehicle standing or parked where it shouldn’t be, humans are responsible for interpreting the nuances of the scene to determine the perceived security risk and the appropriate response. AI assists human operators, helping them to work smarter and faster, but doesn’t replace them.

At Cobalt, we believe security robots should function similarly, making use of trained human operators to make judgment calls about the anomalous conditions that a robot identifies. Is there an intruder in your building, or just an employee working late? Has someone accidentally left a door ajar, or is a non-credentialed individual trying to gain access to secure areas? Human-in-the-loop control allows for further investigation and thorough assessment of situations before law enforcement is alerted or additional resources are unnecessarily deployed. And, when a real threat is verified, it facilitates clear, human-to-human communication between the robot operator and first-responders, providing not just data but intelligent insights and situational analysis that are necessary to be most effective.

By letting robots and humans work together to each do what they do best, you’ll end up with a solution that keeps everyone safer and a security team that’s more effective. That’s why, at Cobalt, we see human-in-the-loop control as the future of security robots and the makings of a pretty perfect human/robot partnership.