| With the gradual development of interconnection,automation,sharing and electricity(CASE)and the improvement of infrastructure and industrial ecology,the era of a new generation of intelligent and connected vehicles is coming.Automotive applications such as advanced driver assistance systems(ADAS),adaptive cruise control(ACC)systems,and smart cockpits continue to emerge,and traditional in-vehicle networks(Controller Area Network(CAN),Flex Ray,etc.)Due to the high bandwidth pressure and high reliability requirements,the vehicle network architecture needs to be changed urgently.With the development of in-vehicle Ethernet and Time Sensitive Network(TSN),the automotive industry has gradually formed a new Electronic/Engineering(E/E)architecture consensus: based on Domain Controller Unit(DCU))of the backbone Ethernet E/E architecture.The new generation of intelligent networked vehicle network architecture will use TSN as the backbone network,and divide communication sub-domains according to functions.Traditional vehicle networks(CAN,Flex Ray,etc.)are used in each functional sub-domain(such as powertrain domain,chassis domain,etc.).The network is connected with other functional subdomains through gateways.The new-generation in-vehicle network architecture can not only maintain the domain division characteristics of the existing distributed integrated architecture,but also meet the requirements of high-bandwidth,low-latency,and high reliability for in-vehicle networks of new-generation automotive applications.The main work and contributions of this paper are as follows:Based on the new generation of intelligent networked vehicle network architecture,this paper completes the design of the CAN-TSN heterogeneous network prototype system,and the prototype system is designed around NXP’s automotive-grade TSN switch chip SJA1105 TEL.The prototype system implements 802.1AS,802.1Qav,802.1Qbv and other in-vehicle TSN related protocols,and on this basis,CAN and TSN are interconnected.This system is an attempt to apply TSN to the in-vehicle network,which can be easily developed for the second time.It can be used for pre-research of the application under the new-generation intelligent networkconnected in-vehicle network architecture.It is important for the application of TSN to the in-vehicle network.meaning.In addition,this paper also proposes a CAN-TSN gateway design scheme based on embedded Linux,the gateway uses a higher frequency(800Mhz)main control chip(I.MX6ULL),so it has better performance.This paper also designs the CAN-TSN gateway structure and scheduling strategy within the gateway.In this paper,a discrete channel gateway structure design is proposed,in which the gateway has good flexibility and expansibility.This paper proposes a new scheduling strategy based on MAT(Maximum Waiting Time)in the gateway,which is based on the CAN-TSN scheduling model,and considers the CAN frame deadline and CAN ID,so it can ensure that the gateway has better real-time.Experiments on the prototype system can prove that: compared with the traditional scheduling strategy(FIFO-based scheduling strategy and CAN ID-based scheduling strategy),this scheduling strategy can ensure that CAN frames have a higher reception rate,using this scheduling strategy The gateway has better real-time performance. |
