Firmware development for an automotive powertrain control item

Environmental concern, fuel efficiency and new market trends push automotive companies, governments and research institutes to explore environmentally-friendly, efficient and sustainable personal and public transportation solutions. Automotive industry is following this trend developing new eco-friendly, smart, and connected vehicles. In this scenario, powertrains are one of vehicle subsystems undergoing several changes in architecture and implementation in favor of electric and hybrid solutions. These innovations within the automotive domain are driven by embedded systems and software solutions. It can be observed that the costs for embedded solutions in vehicles are growing rapidly while mechanical engineering based solutions are stagnating in importance. The motivations of this thesis work are inserted in context related to the new trends of the automotive industry. To be more precise, the purpose of this work was to develop the low-level software for an Electronic Control Unit (ECU) responsible to control an electric powertrain. First, a review of the theoretical aspects of the features to be implemented and required by the system was made to provide a background knowledge about the discussed topics. Then, before to start developing the firmware, a software architecture was carefully defined through multiple software modules, also called software components, with well-defined interfaces. This approach was needed to hide the implementation details of the low-level software and increase the software modularity. The development process, aiming to verify the system requirements, was carried on through a bare-metal approach without any real-time operating system and using the C programming language. Once a first version of the firmware was ready, the testing and verification process was performed using common laboratory instrumentation and debug and trace tools provided by Lauterbach to interact with the microcontroller. Test programs and procedures were developed for each software component to verify the required features trying to maintain the independence with respect to their specific software implementations. Then, the obtained results were analyzed showing correspondences with the expected system responses. As last activities, the integration of the application software, obtained through automatic code generation tools, was performed. Furthermore, a processor-in-the-loop simulation of the motor control algorithm, implemented as a periodic task in the application software, was executed to analyze its worst case execution time and demonstrate the feasibility of its implementation on the target platform. In conclusion, final considerations and further improvements were pointed out.