Signal Packing For CAN FD With Network Security Constraint

The intelligently connected vehicle is an important part of the internet-of-things,and it is considered to be a new type of intelligent terminal device after the smartphone.With the rapid development of the "new automobile four modernizations",the network interaction interfaces and components of the intelligently connected vehicle increased,and the design of the automobile system become more complicated.In this context,attacks against the in-vehicle network often happen,and it makes the security of the in-vehicle network becoming the focus of industry and academia.The Controller Area Network with Flexible Data(CAN FD)is the second generation of the in-vehicle Controller Area Network(CAN)protocol.Compared with the CAN,it has a higher transmission rate and larger transmission bandwidth.In-vehicle network security is an important part of vehicle safety,but the CAN FD has no security mechanisms,which makes the intelligently connected vehicle equipped with the CAN FD very vulnerable to malicious network attacks.Therefore,it is necessary to design a reasonable and efficient security mechanism for the CAN FD to resist attacks.In addition,as a real-time distributed system,the electronic function of the intelligently connected vehicle needs to meet the strict constraints of end-to-end delay,otherwise,the electronic functions of the vehicle cannot run properly.Given the above problems,on the premise of analyzing the common network attack modes,this paper proposed a network security enhancement mechanism for CAN FD named Communication Security Level(CSL)to prevent common in-vehicle network security attacks.Compared with related studies,CSL has a lower bandwidth overhead.Based on the CSL,a greedy-based heuristic signal packing algorithm(GH)and a step-by-step mixed integer linear programming with an initial solution(MILP-IS)are proposed for the different scales input respectively,and both of them are used to optimize the end-to-end delay of the CAN FD.Experimental results show that the proposed algorithms can effectively reduce the average end-to-end delay of the system.As for the average sum of WCRT(Worst Case Response Time),among our experiments,the reduction range of the GH algorithm is[17.88% 51.14%],and the highest one is 90.82%.The MILP-IS algorithm can achieve an average decrease range of [20.05% 61.79%],and the highest one is 90.94%.