USE OF A NEURAL NETWORK ALGORITHM IN MULTIPLE-SENSOR WEIGH-IN-MOTION

This paper describes an algorithm developed to improve the accuracy of Multiple-Sensor Weigh-in-Motion systems. A multilayer feedforward Neural Network is used to combine the output from the individual sensors into one improved estimate of static gross or axle weight. Simulated traffic data indicated that the neural network was largely unaffected by the introduction of a bias in the sensor readings and that different training sets did not influence the results greatly. The addition of noise caused a slight loss of accuracy. Experimental data from a field trial in France showed that the neural network can produce results with a higher degree of accuracy than the simple average method.