Utilização de acelerômetros em uma rede de sensores sem fios de alta qualidade para monitoramento de trens

Abstract

Quality attributes such as speed, reliability, safety and convenience are the essential and decisive factors for the attractiveness of the services provided in urban trains. Meeting these criteria requires an emphasis on rail automation and optimal utilization of available resources. In way to obtain success in automation, real-time information is required from all trains in order to control them efficiently and safely. This work aims at the definition of methods and techniques of highly precise and low cost measurement of acceleration using accelerometers in a ZigBee network for urban trains, and, the evaluation of the information of these devices as a way to validate the possibility of creation of classification models. These techniques should provide an improvement and finalization of the RailBee project. The RailBee consists of a WSN based on the IEEE 802.15.4 standard, composed of embedded remote modules in the vehicles, fixed routers arranged in the roads and base modules for reception of the data. The use of an accelerometer allows the monitoring and evaluation of several measures, among them: speed, acceleration and position, besides other estimates such as: energy efficiency, driver profiles and other techniques that help predictive maintenance. Monitoring takes place in real-time over the Internet and covers a large number of vehicles traveling on permanent roads, increasing traffic efficiency and safety and reducing operating costs. The technology used provides that several quality attributes are ensured, such as: high availability, security in data transmission through standardized protocols, modularity of the architecture and easiness of maintenance. In addition, since this project presents a greater autonomy, greater precision and lower cost of both implantation and maintenance, it emerges as an innovative alternative to the current methods used in the world for urban trains monitoring, generally based on GPS, transmission by satellite or by cellular network. As a means of proving the advantages of such technology, IA techniques and algorithms were applied to demonstrate the feasibility of creating a model to classify instances of values obtained from the accelerometers and gyroscopes.