The Brazilian Traffic Code does not require car owners to install lights in the back of their vehicles, known as brake lights, which light up when the driver steps on the brake. But they are increasingly found in Brazil, whether as factory equipment or installed by consumers who want a little more safety on the road. Four foreign brands and fourteen domestic manufacturers offer this product, including Racetronix Engenharia, based in the city of Bauru, São Paulo. By mid 2007 the company had successfully designed a new system to manufacture brake lights with LEDs (light emitting diodes) by using semiconducting material that lights up on receiving an electrical charge – just as in artificial neural networks and SMD (surface mounted device) components; this feat was only possible thanks to funding granted by Pipe, FAPESP’s Innovative Research Program for Small and Very Small Companies. Since the Racetronix brake light was launched, some four thousand units have been sold in auto shops, according to technologist Antonio Vanderlei Ortega, director of Racetronix.
LED brake lights aren’t really a novelty. They have been around in Brazil since the early 90s. The innovation Racetronix has introduced is the use of artificial neural networks and the industrial assembly of the devices using SMD technology. These systems helped Ortega solve issues related to the intensity of the light emitted by LEDs and the number of diodes that are installed in the equipment. Artificial neural networks are no more than mathematical models, generally implemented in a computational manner, where the operational logic of each part of the process is much like that of our brain’s neurons, interconnected by nodes, with a variety of connections that process the information and facilitate decision-making.
Ortega used the electronic SMD component because it is different from traditional devices since it does not have terminals to insert holes in the printed circuit plate. Assembly of this device is automated and the welding is done on the same side as that on which the component is placed, i.e. on the surface of the plate. “On the one hand, artificial neural networks map out the variables of the process allowing manufacturers to design various brake light alternatives, within the desired characteristics, and selecting the most cost efficient. On the other hand, the use of SMD components reduces manufacturing costs by letting one eliminate two phases of the electronic assembly process – the pre-molding and cutting of terminals,” states Ortega.
The retail price of Racetronix’s product ranges from R$ 27.00 to R$ 34.00. There are cheaper models, down to R$ 9.00, but not with 24 LEDs as is the equipment Ortega supplies. He guarantees that his product is among the cheapest on the market. Quite often the wholesale price is less than R$ 10, depending on payment terms and purchase quantity. He believes that one must compare this price with the product being purchased. In his case, for instance, this product complies with engineering criteria as well as with the regulations of Contran, the Brazilian Traffic Department. “That is why the best thing is to compare Racetronix’s prices with those of Arteb and Cibie (leading auto light manufacturers). My products, with 24 LEDs, are no better than theirs, but are definitely cheaper.”
It took five years of research to develop this product. The earliest designs were developed in 2002. Ortega previously worked at a multinational that manufactured LEDs, among other products. “I was responsible for LED quality control and was already acquainted with the complex behavior of light emissions.” The idea of mapping these effects originated while he was completing his industrial engineering master’s degree at Paulista State University (Unesp) in the city of Bauru. That was when his counselor, Professor Ivan Nunes da Silva, introduced artificial neural networks to Mr. Ortega. “We researched scientific articles and didn’t find anything about the use of neural networks to manufacture brake lights,” recalls Ortega. While doing his doctorate, he expanded the technique in order to apply it to break lights using LEDs. Given its complexity, neural networks were not sufficient to map all variables involved and that is why there was a need to use another system, called fuzzy logic, or fuzzy inference systems. This theory, developed in the 60s, sought to establish a technique that allows one to process imprecise or vague data.
FAPESP’s Pipe project was created between his master’s degree and doctoral studies. “At first, it helped to improve and put into practice what was developed during my master’s degree studies, such as testing other network models, using SMD components and developing a complete product. In the following stage, equipment acquired during the project was used to generate samples used in the doctorate, to measure light intensity and conduct quality control testing.”
The brake light’s key function is to increase the vehicle’s safety by reducing the risk of other drivers crashing into the back of the vehicle. A study carried out in 1998 by the Department of Motor Vehicles in the US (where the use of the device in passenger cars has been mandatory since 1986) concluded that brake lights help avoid 92,000 to 137,000 collisions a year in the US. Out of this total, between 58,000 and 70,000 would injure the vehicle’s passengers or drivers. According to the study, by avoiding these accidents the country is able to save some US$ 655 million a year in property damages.
On route to auto manufacturers
Although Racetronix’s brake light is specifically targeted to the auto parts replacement segment, the company’s medium-term goal is to supply automakers directly. Soon Ortega also intends to introduce two other novelties in the market: a brake light with a reverse light and LED tail lights. The first will be marketed to auto parts resellers, but the tail lights are likely to attract automakers in Brazil, in line with the trend found among certain imported models that already have LED tail lights rather than conventional incandescent lamps. “Since one must have a strict quality control system (like ISO 9000 or QS 9000) to supply automakers, our next step is to implement one of these systems and get certification,” concludes Ortega.
Manufacture of brake lights using artificial neural networks and SMD components (nº 04/08973-9); Type Pipe – Innovative Research Program for Small and Very Small Companies; Coordinator Antônio Vanderlei Ortega – Racetronix; Investment R$ 97,741.86 (FAPESP)