• 
• 

Published on August 14, 2007

Equipped with an intelligent system, this electric vehicle can drive itself. It can decide to move forwards, turn left or right, or stop on its own.

The intelligent car is a result of efforts by a group of engineering students formed under the Jack-O-Lantern team from Chulalongkorn University. Their aim is to make a new kind of vehicle that can drive itself automatically.

As the car is designed to be much smarter and make its own decisions with no need for human control, it could be used for people with disabilities who could not drive for themselves, Kamol Chuengsatiansup, one of the development team, said.

The team spent six months developing an intelligent system for the car and it recently won first prize in the Thailand Intelligent Vehicle Challenge organised by the Thai Robotics Society and Asian Institute of Technology, sponsored by Seagate Technology (Thailand).

Kamol said the intelligent system was the heart of the car.

The team adopted artificial intelligence (AI) technology to develop software to analyse information for the car's decision-making.

Input information is vital for data analysis so the team integrated laser scan, image processing and global positioning system (GPS) technology to get information from all around to produce more accurate results.

Two digital cameras sit on top of the car to capture wide-angle images while a laser scanner lower down detects obstacles within a 180-degree angle and 30 metres in distance. Meanwhile, GPS and a digital compass are used to identify the car's current position and calculate the direction for movement.

Chanin Chanma, another developer, said all information from the camera, laser scanner, GPS and digital compass is sent to the central system for further processing.

The AI software determines the direction the car should move.

"We analyse data from the images we receive to see the overall environment, but image information from the cameras alone is not enough to make accurate decisions so we also rely on information from the laser-scan sensor as well as GPS to make the software produce accurate results," he said.

Meanwhile, to make the car decide to stop, turn left or turn right, the team also integrated image-recognition software so that once the camera captures an image such as a traffic sign, the system knows what the sign means and makes the decision for the car.

Chanin said that after the system processed all the information, it gave a result on how fast and in what direction the car should go and this information was sent to a microcontroller to control the car's mechanical systems such as steering, accelerator and brake pedal.

As the car is now just a prototype, its top speed is limited to 10 kilometres per hour.

However, Kamol said the car was just the first version and the team planned further development to improve the car's decision-making system.

"We also hope the system we develop will eventually be used in commercial cars as the intelligent system will assist safer driving," he said.

Pongpen Sutharoj

The Nation