A large share of innovation and customer value in modern vehicles comes from advanced computer-controlled functionalities. With the increasing volume of such functionalities, the vehicle software has tremendously increased in size and complexity over the past few years. Already today, the software in a modern car consists of millions of lines of code that runs on tens of distributed Electronic Control Units (ECUs) that can be connected by various types of in-vehicle networks, such as CAN, Flexray, LIN and Switched Ethernet. The safety-critical nature of several vehicle functions puts real-time requirements on them. The developers of these functions are required to ensure their predictable timing behavior.

In addition, the advanced features in modern vehicles require new levels of computational power (e.g., due to more data-intensive sensors, such as video camera, radar and ultrasonic sensors) and more complex coordination among subsystems (e.g., engine torque and brake force for stability control). Multi-core ECUs offer a promising solution for running such computation-intensive vehicle functions. However, multi-core ECUs are more prone to unpredictable behavior (due to share caches and memory banks) as compared to traditional single-core platforms.Clearly, there are many challenges related to handling the complexity, allowing reuse, supporting the required predictability, verifying timing requirements, and dealing with mixed criticality during the development of vehicular embedded software systems. The objective of this Special Session is to provide a forum to the researchers and practitionersfor discussing novel and emerging solutions to these challenges.