Open-VSeSeMe: A Middleware for Efficient Vehicular Sensor Processing

The recent increase in the number of sensors within cars has resulted in various fragmented software stacks and development frameworks. In this ecosystem, applications have to make sense of raw sensor data themselves. As a remedial solution, we present Open-VSeSeMe, a middleware atop TinyOS that converts raw sensor streams into data units with semantic meaning. These data units can be shared between applications leading to efficient use of resources. In addition, we argue that the use of a common software stack leads to hardware standardization and opens up the platform to third-party developers, making a Car App Store possible. Furthermore, the entire architecture is event-driven which frees the applications from the clutches of constant polling. Finally, using a number of illustrative examples we show the utility and usefulness of Open-VSeSeMe.