Renault Sport, whose engines power the blue and white Rothmans Williams Renault cars to be driven by Damon Hill and Nigel Mansell in the last three races that mark the climax of this season’s Formula One world championship, intends to maintain its technical advantage in Formula One motor racing by keeping one step ahead in the precise task of ensuring a car is sufficiently prepared for a race, 01 Informatique reports. There are something like 30 different parameters that must be constantly checked and analysed on every car, from both the chassis and the engine, before each race and in real time during the race. The slightest malfunction could result in disaster both for the team and the driver. Engineers in the pits send the driver information by radio which is displayed on the windscreen. There are also alarms that are set off when those factors reach a critical threshold, such as when the temperature of the engine goes above its safe operational level. Another problem is that there is very little time during testing to correct any problems that may crop up.
Telemetry
In 1993, Renault Sport and Compagnie des Machines Bull SA, already partners on the telemetry application, decided to collaborate on an improved system that can analyse the relationship between the factors and warns the engineers if a problem is discovered, using advanced computing techniques such as artificial intelligence. The new system gives the engineers data on which they can make predictions for the coming race during testing and intervene intelligently at points during the race. The system recognises, for example, the effects of temperatures that can be produced when the driver goes too close to another car, which can hamper the engine’s ventilation. The expertese of the engineers on the cars is collected by Bull Cediag computer scientists using the Open Kads data collection and information modelling tool, which picks their brains and is capable of understanding a generic model dedicated to surveillance and alarm applications. This tool generates a C++ program to carry out the monitoring in real time. Finally the graphical interface, developed using Ilog SA’s Ilog Views graphical interface design library and running on a Silicon Graphics Inc workstation in real time, displays the telemetry data collected by the system. The new system achieves three main goals: firstly it monitors the level of performance, determined by the behaviour of the engine, which is expressed through contextual thresholds and complex calculations on the telemetry data; secondly, it detects malfunctions at an early stage; and thirdly, it filters out unneccessary alarms. Knowledge about the normal functional limits of the engine can be compared with the real limitations experienced, enabling malfunctions of the engine to be predicted faster than before. The combination of this theoretical knowledge and the expertese the driver has about the way the car performs on the race track enables alarms to be filtered so that only the useful ones are given to the Renault engineers. The system can for example recognise situations that are the direct result of the driver’s action, such as braking.
French Grand Prix
The system took one-man year to develop and has been in use since the French Grand Prix in July. The basic development knowledge base is brought back up to date before each new race, and from one race to the next the Renault engineers spend 12 days working with Bull computer experts improving the system’s expertese, by adapting it to different environments being encountered, such as the track and weather conditions or technical modifications that have been made to the vehicle.
