Dubai Police deploying self-driving AI rovers with onboard drones

While Waymo and Cruise are still grinding away on robotaxis in the US, Dubai is taking autonomous driving to neighborhood watch duty.

Dubai Police just revealed a fully electric, driverless patrol vehicle designed specifically to roam residential areas, with testing expected to wrap ahead of a rollout next year.

Slow rolls, sideways moves, and pocket drones

Do not expect any Michael Bay car chases out of this machine. The robocar maxes out at a leisurely 4.35 mph (7 kph), basically brisk walking pace.

Speed is not the point here. Built by Micropolis Robotics, the vehicle was developed by 65 engineers over five years and features a completely custom mechanical platform.

Each wheel has independent steering, braking, and throttle control. That allows it to drive forward, backward, slide sideways, or even pull off donut spins on the spot. It also moves almost silently.

Under the hood, the battery packs enough juice for 15 hours of continuous patrol. The surveillance rig includes a 360-degree camera setup with facial recognition, paired with an AI engine trained to detect suspicious behavior.

If it needs eyes in the sky, it carries its own camera-equipped drone that can launch directly from the vehicle to snap aerial footage.

When the AI flags something fishy, the car will not intervene directly. Instead, it beams an alert back to the Command and Control Center at Dubai Police HQ, leaving the actual response to human officers.

High-tech toys are standard operating procedure in Dubai. The department already fields supercar squads, previously deployed an informational robocop, and even handed firefighters water-powered jet packs.

Why it matters

Instead of tackling complex 60 mph highway navigation, this vehicle zeroes in on a much smarter, safer deployment zone: hyper-low-speed perimeter surveillance in suburban streets.

Capping the speed at walking pace effectively eliminates most catastrophic crash risks, while the 15-hour battery life solves the grueling chore of round-the-clock neighborhood patrols. Pairing ground observation with a deployable drone covers visibility gaps cleverer than piling more sensors onto the roof.

That said, rolling facial recognition and automated suspicion detection straight into residential front yards raises immediate false-positive and privacy concerns. If the system flags regular citizens taking late-night trash runs, it risks overwhelming human dispatchers with useless alerts.

Still, as a practical template for city-level autonomous robotics, shifting the burden from pursuit to quiet, persistent observation seems like the right move.