Time-triggered Ethernet Real-time Analysis And Switching Mechanism Design For Connected Vehicles

In connected vehicles,intelligent systems such as the Advanced Driving Assistance System and Adaptive Cruise Control System have led to a significant increase in in-vehicle data.Current in-vehicle buses cannot meet the transmit demand of high bandwidth and low delay.Time-triggered Ethernet is a candidate for the next-generation connected vehicle’s backbone network because of its advantages in real-time,bandwidth,and determinism.For safety-critical vehicle systems,it is necessary to have the shortest possible transmission delays in terms of in-vehicle communication.The correct behavior of the system depends not only on the correct transmission of messages but also on the ability of the system to respond within a defined period after receiving a message.On the other hand,as the hub of in-vehicle information transmission,the switch takes on the critical responsibility of the whole vehicle information interaction.The efficiency of data exchange and transmission in the switch directly affects the reliability and real-time performance of the automotive system.Therefore it is essential to have an accurate and efficient time-triggered approach to real-time analysis in Ethernet and a practical design of switching mechanisms.The main work accomplished in this paper is as follows:This paper proposes a holistic delay analysis approach with high accuracy and low time complexity.We propose a new approach based on the holistic method because the existing realtime analysis approaches do not balance calculation results’ accuracy and time complexity.The approach obtains an upper bound on the end-to-end delay by recursively computing the delay of a frame in all network nodes it passes.We use the pre-defined scheduled instants in the TT table as input parameters in the analysis process.Then we consider the impact of the offset of the data flow on the delayed results.On the one hand,this approach reduces the pessimism brought by TT traffic on the analysis results.On the Other hand,it reduces the computational expenses caused by loops traversing unnecessary data flows.The experimental results show that the approach proposed in this paper reduces the time complexity of the analysis while improving the accuracy of the calculation results.Compared with current methods with higher accuracy,the computational overhead is reduced by about 90%,which is more suitable for real-time analysis of medium and large configuration networks.In this paper,a high-priority switching mechanism is designed.The switch is responsible for in-vehicle communication,and the efficiency of its data exchange directly affects the reliability and real-time performance of the automotive system.Domestic research on time-triggered Ethernet is at the theoretical stage,and no mature commercial product has been developed yet.In this paper,we design a multi-priority switching mechanism.The switching mechanism supports two priority levels of RC traffic,and the high-priority RC traffic is transmitted in the switch queue first.In addition,different output ports correspond to different output units,so frames from different ports can be transmitted simultaneously,reducing the delay of frames in the switch.Finally,the delay of multi-priority RC traffic in the switch is analyzed by the algorithm and compared with the delay of fixed-priority RC traffic.To verify the effectiveness of the proposed real-time analysis method,we build a simulation model of time-triggered Ethernet in this paper.The node model of the network is established using the OMNET++ simulation platform.We invoke the model library and the extensions of the existing library functions to specifically implement the logic inside the nodes and complete the implementation of the simulation model.Finally,simulation experiments of end-to-end delay are made using the model.We compare the results with the analysis results of the approach proposed in this paper,which verifies the correctness of the theoretical analysis approach.The simulation model can also be used to evaluate the system’s performance and provide a basis for developers to design and improve the system configuration.