Over the past two decades the manufacturers of internal-combustion engines that are used in motor vehicles have been very successful in reducing the harmful side effects of their products on the environment. However, they are under ever-increasing pressure to achieve further reductions in the quantities of polluting gases emitted by the engine, and a decrease in the amount of fuel consumed per kilometer. At the same time, vehicle characteristics that are desirable to the driver must not be compromised. Satisfying these diverse requirements requires precise engine control and comprehensive monitoring of the operational parameters of the power unit. Engines are highly price sensitive, and it is desirable to achieve the increased level of measurement that is required for enhanced control without additional sensory devices. Thus, the indirect estimation of quantities of interest using virtual-sensor techniques, without direct measurement using dedicated sensors, is a research area with considerable potential. Intelligent-systems techniques, such as neural networks, are attractive for application in this area because of their capabilities in pattern recognition, signal analysis and interpretation. For this reason, the use of neural networks in the monitoring and control of motor vehicle engines is an area of research which is receiving increasing attention from both the academic and commercial research communities. A virtual-sensor technique, the Virtual Lambda Sensor, is described here which uses a neural network for the estimation of air-fuel ratio in the engine.