بکارگیری سیستم های هوشمند برای مشخصات تجهیزات خودرو با استفاده از LED ها

The advantages offered by the electronic component light emitting diode (LED) have caused a quick and wide application of this device in replacement of incandescent lights. However, in its combined application, the relationship between the design variables and the desired effect or result is very complex and it becomes difficult to model by conventional techniques. This work consists of the development of a technique, through artificial neural networks, to make possible to obtain the luminous intensity values of brake lights using LEDs from design data.

The LED device is an electronic semiconductor component that emits light. At present time, it has been used in replacement of incandescent lights because of its advantages, such as longer useful life (around 100,000 h), larger mechanic resistance to vibrations, lesser heating, lower electric current consumption and high fidelity of the emitted light color [1]. However, in designs where incandescent lights are replaced by LEDs, some of their important characteristics must be considered, such as direct current, reverse current, vision angle and luminous intensity. In automobile industry, incandescent lights have been replaced by LEDs in the brake lights, which are a third light of brakes [2]. In these brake lights are used sets of LEDs usually organized in a straight line. The approval of brake lights prototypes is made through measurements of luminous intensity in different angles, and the minimum value of luminous intensity for each angle is defined according to the application [3]. The main difficulty found in the development of brake lights is in finding the existent relationship between the following parameters: luminous intensity (IV) of the LED, distance between LEDs (d) and number of LEDs (n), with the desired effect or result, i.e., there is a complexity in making a model by conventional techniques of modeling, which are capable to identify properly the relationship between such variables. The prototype designs of brake lights have been made through trials and errors, causing increasing costs of implementation due to time spent in this stage. Moreover, the prototype approved from this system can sometimes not represent the best relationship cost/benefit, since few variations are obtained from configurations of approved prototypes. The artificial neural networks are applied in cases like this one, where the traditional mathematic modeling becomes complex due to nonlinear characteristic of the system. These networks are able to learn from their environment and to generalize solutions, making them attractive to this type of application.

This work presents a technique based on use of artificial neural networks for determination of luminous intensity values for brake lights (IVBL), in which are considered the main designs characteristics. Therefore, the developed tool constitutes a new technique that can efficiently be applied in this type of problem. The developed methodology can also be generalized and used in other applications that use groups of LEDs, such as in traffic lights and electronic panels of messages. The developed tool has significantly contributed for reduction of costs in relation to implementation stage of brake lights, i.e., it minimizes spent time in prototype designs. The tool has also allowed simulating many options for configurations of brake lights, providing the choice of a sample that offers an appropriate relationship between cost and benefit.