The UK Neural Computing Applications Forum recently held another of its updates on the the latest in neural computing research and its practical applications, this time at King’s College, London. Cutting through the complex mathematical formulae beyond the abilities of the average human neural capacity, the practical uses and research areas prove ever more obscure and imaginative. Lucas Advanced Engineering of Solihull is working on the Driver Support System, a research programme started four years ago in conjunction with Jaguar Cars Ltd and Southampton and Oxford Universities, to improve road safety and traffic efficiency. The system effectively acts as a co-pilot to the driver and has two main functions. Firstly it acts as a warning system if the driver deviates from standard driving practices and secondly it develops an awareness of the road system around you, such as the distances between you and the car in front, beside and behind. The neural network is trained to recognise normal driving behaviour of the driver and if he gets closer to the car in front for his speed than normal or drifts to the side of the road, if for example he is falling asleep, or an obstacle appears ahead suddenly then a warning loudspeaker and visual LED driver display warn of the change and alert him to the danger. The equipment includes an integrated radar sensor at the front of the car and a processing rack at the back giving information within at least the normal safe stopping distances. Sensors on the side of the car can warn of obstacles in the blind spot to aid safe overtaking or the feasibility of overtaking if an obstacle suddenly appears in front. Another aspect of this project is under way to combine this device with an assessment of the weather conditions determined by information such as external temperature and daylight availability, so that practical safety stopping distances can be determined for each specific situation. A prototype model is should to be ready this October.
