Memory Measurement and Message Usage Improvement on an Elevator Embedded System

Abstract: All embedded systems are unique, a design that is suitable for one system can rarely be copied over to another. This inherently makes designing embedded systems difficult. The difficulty is only amplified by the uncertainty of the future requirements as it is developed over time. Being able to continuously validate the performance and the reliability is of great importance to be able to ensure fault proof execution.This thesis explores two areas. A method of tracking the static and dynamic memory usage of a system is crucial to ensure correct functionality under all conditions, and that the implemented hardware will suffice. Multiple possible tools, each functioning uniquely, were developed and tested to find the most suitable for measuring the memory usage of the elevator system. Additionally the message usage, i.e. the way the different units within the studied system communicate with each other, was scrutinized for possible performance and reliability enhancements. A study was made for the most optimal communication protocol, and for how the transmissions could be improved upon.The results show that for this specific system, the best way of calculating the memory usage is with a tool developed within this project. Using this tool it was found that none of the modules in the elevator system use more than 30 % of the available memory during execution. The message usage study shows the most optimal protocol is CAN with the ISO 15765-2 upperlevel protocol, which is the one currently in use. However, improvements to the message transmissions are suggested, such as taking full advantage of the CAN protocol and by implementing message buffers on the receiving end.An important conclusion is that just as there is no unique system design that fits all, there is no memory measurement tool or message usage implementation that fits all systems. Each system has to be analyzed to find the most optimal solution for that particular system.