According to the team, with a single detector their technique can track an object 200 times faster than traditional video-based methods, without any prior information and with minimal computing power.