Collision-prevention technology

A new technology will use radar, lasers and camera sensors to help prevent accidents in built-up areas

The number of pedestrians killed in urban traffic accidents could be reduced with new collision-prevention technology that automatically slows a car down when a person moves into its path.



Researchers based at

Warwick University

are working on developing such advanced driver assistance systems (ADAS) with support from Jaguar Land Rover. The team, led by Ying Ping Huang and Ken Young from

Warwick Manufacturing Group

, hope to have a proof of concept demonstration vehicle in three years.



Their technology would enable cars to use radar, lasers and camera sensors connected to a central control unit to detect and classify potential obstacles in city environments. The radar and lasers, mounted on the front of the car, would send radio waves and laser pulses respectively to determine the distance of potential obstacles. The range to the object would be determined by measuring the time delay between transmission of a wave or pulse and detection of the reflected signal.



Two camera sensors, installed on top of the windshield behind the rear-view window, would gather 3D images of the potential obstacle, a technique commonly used in mobile robotics. The cameras take two pictures of the same scene and computer software compares the images. The program uses the disparity (the amount the images shift) to calculate the object's distance.



Ken Young said that with the use of algorithms and motion-analysis software the object can be classified as static or in motion.



If the central-control unit determines the situation is potentially dangerous, it signals to the engine or braking system to decelerate.



This technology is used in cars with adaptive cruise control, which uses radar to gauge the distance of a car in front and automatically adjusts speed to maintain the correct space between the vehicles.



However, most current automotive collision-prevention technology is designed for motorway use. Young said motorways are 'fairly controlled environments' and designing anti-collision technology that can work in an unpredictable and dynamic setting such as a city is much more difficult.



The most challenging obstacles in urban environments are pedestrians. Young said their safety has been overlooked by automobile manufacturers for many years. 'Obviously over the years vehicles have become pretty safe for the drivers and their occupants, but one of the things we haven't made much progress in is pedestrian safety,' he added.



According to the European Road Safety Observatory, of all traffic fatalities in EU countries, the proportion of pedestrian fatalities is about 17 per cent. The EU has introduced legislation that it hopes will cut the number of pedestrian traffic deaths in half by 2010.



Car companies have also made attempts to increase pedestrian safety in recent years. In January, Volvo Cars introduced a concept car with built-in pedestrian-collision avoidance technology at the North American International Auto Show in Detroit.



The Volvo S60, which will be launched in 2010, uses a wide-angle radar to spot obstacles at the side of the road and a windshield-mounted camera sensor to attain an image of them. If the car's internal image-analysis software identifies an object as a person, the camera tracks the image.



Young's research team at Warwick is working on a similar concept to Volvo's but their technology will rely less on radar. He said that the use of two cameras will provide more information about the shape and movement of an object.



The challenge will be to ascertain how to process this additional information. 'We've got to be able to do it in real time so we need fairly simple algorithms to run on today's processors,' added Young. 'My suspicion is a lot of what we come up with may not be immediately useable, but it may well be useable in a few years' time as processing speeds go up. We can't limit ourselves to using today's hardware.'



The autonomous trend was recently demonstrated by Jaguar Land Rover. On March 3 the company unveiled a hybrid vehicle that uses GPS and mapping data to automatically adjust vehicle speed depending on local speed limits, traffic conditions and road features such as bends. It is also programmed to decelerate when travelling past schools, approaching traffic lights and pedestrian crossings. The vehicle, using control technology dubbed Sentience, will be commercially available in Jaguar Land Rover cars in four years.



Young and Huang's ADAS may be available in commercial road vehicles in around a decade.



'I think pedestrian safety is something where legislation is going to demand that vehicle manufacturers do something about it,' added Young.



'We're just trying to get ahead of the game.'



Siobhan Wagner