The Study and Implementation of Efficient ADAS Concurrency Task Based on Some/IP

By stating the key essential technique urgently needs to be breakthrough of super complex system architecture in the developing strategy of smart vehicle, bringing that the Advanced Driving Assistance System (ADAS) is the best instance to reach this developing strategy target. Abstracting a computer system model from ADAS, bring up the key basic technique it is to solve ADAS efficient concurrency task in complex system based on Scalable service oriented middleware over internet protocol (SOME/IP), which means “Rapid Controller area network (CAN) data collection” and “Dispatching CAN instruction on time”. Simplifying the super complex system architecture of vehicle to be as a small Automotive Ethernet (AE) model, by using Common application programming interface (CommonAPI) to design and implement the Application Programming Interface (API) of concurrency task under Service Oriented Architecture (SOA) frame thinking, coding to implement API and verifying functions of concurrency task, by through result of test program to evaluate API executing performance. Theoretically, the API designed by “CAN signal oriented” thinking brings massive context interaction just as going in and out of AE stack, which takes more AE bandwidth too. Optimizing the API of concurrency task by using “CAN signal set service” thinking, programming and testing again, the test result shows that the optimized API descends the massive context interaction and obviously improves AE bandwidth usage. By using “theory guiding practice” this way, at last this paper successfully implements efficient concurrency task this study goal.