The ability of computers to distinguish objects is continually improving, but they tend to operate in isolation. What's the use of a security robot if it can't share information about the attacker with other machines? Perhaps scientists at Cornell University have found a solution to this problem. They are developing a system that would allow robots and other autonomous devices to cooperate with each other to identify and track objects. Each bot will send its images to the central unit, which can communicate both with other systems and with online cameras to speed up the process of object identification. The group will come together to compare images of objects, monitor them and collect much more information than if they did all the work on their own.
For example, a robot can detect a suspicious person sneaking through a crowd. A team of robots can track all of his movements. Using the planning function, the system will determine which cameras need to be turned on and which robots have to be used in order to get the best view.
Scientists still have a lot of work ahead of them. It is necessary to complete and test the system, not to mention preparation for commissioning. However, these efforts are likely to be much more beneficial than strengthening security measures in other ways. The United States Office of Naval Research is supporting the project, and it is hoped that the Navy will use the system under development to coordinate drone operations. Regardless of what it turns out to be, logic dictates that the time of lone wolves is already ending.