Headlight performance – consumer information

　　　　　　　　　　　　　　　　　　　　Kåre Rumar, Sweden

Keywords: Headlight performance, consumer information, comparison, safety, rating

1. Background

Every road crash is a result of a break down of the interaction between driver, vehicle, and road. Some of the characteristics of the car the driver learns quickly. For instance, he will soon notice if brakes or stability are bad. But other characteristics are very difficult for the driver to notice. Such characteristics are for instance crash performance and headlight performance. Good consumer testing and information of such characteristics can give the driver the crucial information he would otherwise not get.

Driver visual input is accepted to be the major external information source for driver decisions and behaviour. Consequently certain visual performance is required in order to receive a driver’s licence in most countries (acuity, visual field, etc.). However, the vehicle always sets certain limits for human vision and no driver can see better than the vehicle permits (e.g. visual field, headlight performance). These limitations caused by the vehicle are probably less obvious and most pronounced concerning headlight performance.

To a very large extent headlight characteristics decide driver visibility distances. Still there are only some minimum regulatory demands and requirements on the headlights (ECE, SAE, JAE). And most of the driving in darkness is carried out on low beams. Therefore low beam performance is the most crucial part of headlight performance both from safety and driver comfort point of view. The visibility distances offered especially by the low beams are very often too short to be safe (Rumar 2002, Flannagan 2003).

Unfortunately drivers and potential car buyers have very limited ability themselves to estimate, analyze, and rate headlight performance. Such analyses normally require full-scale field comparisons under full experimental control. Experiences show that especially inexperienced drivers have a tendency to rate such low beam headlights high, that offer much light close to the car, while low beams, which offer less light close to the vehicle and more light further down the road are rated lower. This make subjective driver rating very invalid and often even inverted. Studies have shown that drivers systematically believe that their visibility distances in night driving are longer than they really are.

The consumer tests concerning headlight performance (if any) are so far very bad. They are often very vague and unspecific. It is hard to understand what they are rating. The potential buyer gets some information but not very much. There is a need to develop a more extensive and detailed rating system for headlights. The largest car & consumer magazine in Sweden (Vi Bilägare) wanted to fill this gap and took the initiative to this project.

2. Problem

The task of this project was to develop a measuring system, which could offer potential buyers and drivers of automobiles on the market a relevant description of headlight performance with acceptable reliability and validity. The results should be detailed enough to describe headlight performance in such a way that the reader could make his or her own estimation of headlight performance and not only give an overall rating value. The measuring system should be quick, simple and comparatively inexpensive.

The cars should be measured as the consumers get them on the street, when bought by the car sales company new or in used condition. No modification should be made except correct aiming of the headlights in the same way as workshops and car inspection authorities do them. In other words, the results should be representative of automobile headlights as consumers normally get them.

3. Method

The method developed is based on a project in progress at Transportation Research Institute, University of Michigan (UMTRI) (Flannagan and Sivak 2001, Flannagan 2003), earlier efforts to rate headlight performance carried out by car and headlight manufacturers and our own thoughts and experiences. This project is slightly different in that the ambition is to describe headlight performance to the reader in such a way that he can do his own judgement of the headlights based on the driving situations relevant to him.

This project contains several sub-problems. First, how should the lighting characteristics of the headlights be measured? And secondly, what in the light distribution should be measured? Thirdly, how should the measurement results be weighted and rated? And finally fourthly, how should the results be presented to the drivers?

3.1. How to measure the headlight light distribution?

There are at least three different ways to approach the measurement problem. The most straightforward and valid way is to measure low and high beam light distribution in full scale on a straight flat road. This approach is, however, very difficult, time consuming, and costly. Furthermore, it is not possible to use during the Swedish summer, which is very bright. The second approach is to mount the headlights one by one in a lighting laboratory and use a precision goniometer to make the measurements. The results can then be analyzed and summarized so that the total light distribution in front of the car can be calculated. This approach is however less valid because the headlights are dismounted from the cars. It is also quite expensive because a lighting laboratory is required.

Therefore a third approach was chosen. The full-scale situation was scaled down by means of a fresnel lens in a precision built optical box and each headlight on the car was measured separately in a garage (see figure 1). Correct headlight height, which is an important variable, is automatically obtained and included by this method. Design, construction, and validation of this special optical box were carried out by an optical company (Optab 2003), which had considerable experience from advanced optical projects for the Swedish Defence. This approach was chosen because it was judged to offer enough validity and reliability at acceptable costs. According to Optab, valid and reliable measurements can be achieved at distances of at least 1,000 meters in front of the car.

Figure 1: The optical box used to measure the headlight distributions in a garage.

The measurements of the headlight distributions are carried out in 5 to 10 minutes time once the car is correctly aimed longitudinally. A total of 15 million points can be measured. A computer programme immediately draws the iso-lux diagrammes and makes the calculations necessary to specify the various headlight performance characteristics and the points of merit in comparison with the base sample of cars.

3.2. What to measure in the light distribution?

There are also several ways to approach the question what to measure. One way is to use the measurement points used in the ECE and/or the US regulations or the proposed harmonized regulation. This approach however does not give the reader any understanding about how good the headlight performance is.

Therefore another approach was chosen. The scene in front of the driver is divided into a limited number of areas, which are important to illuminate during driving in darkness. This approach was used by car manufacturers and was also used in the UMTRI-project mentioned. Within each one of the chosen areas an average illumination is calculated. This method offers a simple and inexpensive solution of the problem at the same time as it offers acceptable validity and reliability.

The first step was to specify the road scene of interest. The road is straight and flat. Each lane is given the width 3.5 metres. In accordance with the UMTRI proposal the road was divided into three lanes. The car is supposed to move in the centre lane, the left lane is representing the lane for oncoming traffic and the right lane is repre

[1] [2] [3] 下一页