Dubai students develop drones that take spoken commands, driverless race car

· United Arab Emirates

Dubai students develop drones that take spoken commands, driverless race car

Inspecting a railway line could one day begin with a spoken instruction rather than a series of programmed commands. Aditya Shibu, a student at Heriot-Watt University, is developing a drone-control system that aims to let people direct multiple aircraft simply by telling it what they want done.

His project is among student projects that explore how artificial intelligence could change the way people operate machines from drones carrying out inspections to a race car navigating a track without a driver.

For Shibu, the aim is to make drone technology accessible to people who do not have specialist programming skills. Instead of coding flight instructions or setting individual waypoints, an operator would give the system a task in everyday language.

“You can tell these drones, I want to inspect site A. It’ll decide the tasks,” Shibu explained.

The system uses the team’s own AI model to interpret instructions and allocate tasks across the drones, he said. The approach is intended to go beyond the choreographed formations associated with drone shows and support practical inspection work.

One proposed application involves inspecting metro railway infrastructure for Dubai’s Roads and Transport Authority (RTA). Shibu described a proposal combining drones with base stations that communicate with one another to inspect tracks, with potential usefulness during flooding.

“It is a proposal that we are trying to take forward,” he said.

Building inspections are another possible use. An operator could ask the system to inspect a particular part of a building without having to manually programme each aircraft’s movements.

He also referred to work involving Etihad Rail according to Shibu, the current demonstration allows drones to receive instructions and travel to a specified location. Cameras and other equipment could be added to suit an inspection task.

Virtual reality offers another way to interact with the system. Shibu said users could control drones through a VR interface while viewing information such as battery status. Gloves that would allow users to control drones through hand movements are a longer-term ambition.

From a small prototype to the race track

A separate project by Atlas Racing is working towards an autonomous Formula race car, with its driving software already tested on a small prototype.

Unlike a vehicle following a fixed route, the prototype detects cones marking both sides of a track and plans its path as it moves, Anosha team principal explained.

“The purpose of the vehicle is to move within the track, and the path is not predefined,” she said.

The software must identify the track boundaries and work out a route between them. Steering and speed-control systems then help keep the vehicle on course.

The next step is to move towards a full-size vehicle. She said an existing golf cart at the university was used for testing.

Atlas Racing’s competition record provides a foundation for that ambition. The team, listed by the Institution of Mechanical Engineers as representing Heriot-Watt University Dubai, won the FS-AI Real World Award and the Babcock FS-AI Overall Statics Award at Formula Student UK in 2025.

Ras Al Khaimah to test driverless cars in real-world environments for potential useEmirati inventor develops first AI system to warn motorcycle riders of road hazardsDubai to cut traffic congestion by 5%, accidents among new drivers by 20% using AIUAE to have over 10,000 autonomous taxis, delivery vans by 2027, official says

Source: Khaleej Times - UAE