Research On Real-time Scheduling For Static Segment In Flexray Systems Based On Hybrid Genetic Algorithm

With the development of X-by-wire system, high speed transmission, reliability and supporting distributed controlling are asked for in-vehicle network. Most of the current in-vehicle networks are based on the CAN bus which is triggered by events. Since the CAN does not exceed 1Mb/s maximum speed and its bandwidth utilization is low, it is difficult for CAN to obtain the advantages in a large amount of data switching application. Thus CAN is unable to meet the requirement of the fault-tolerance and high bandwidth in next generation X-by-wire system applications. Therefore, BMW, Daimler-Chrysler, Freescale, Bosch and Philips set up FlexRay Union to build the FlexRay standard protocol. FlexRay, the new generation vehicle bus, has the advantages of high speed, dependability and real-time transmission. The FlexRay protocol offers various benefits over the existing in-vehicle network, and it can be successfully used for safety-critical automotive applications such as network-based braking or steering systems.FlexRay protocol is constituted by the static segment and the dynamic segment, while the static segment is based on Time Division Multiple Access (TDMA) to achieve deterministic behavior, the dynamic segment adopts so-called Flexible TDMA (FTDMA) to achieve flexible and efficient use of the network. Therefore, scheduling of the static segment and creating the static scheduling table are critical for an efficient FlexRay-based system design. This paper mainly focuses on the real-timing scheduling for static segment in the FlexRay systems.Based on the analysis of FlexRay protocol and its implementation, the paper designes the application model for the FlexRay static segment scheduling, and this thesis defines the performance metrics—bandwidth utilization, analyzes the static real-time scheduling constraints in detail, focusing on the response time constraints to ensure that the messages can be successfully transmitted before their deadlines. Scheduling for static segment in FlexRay systems is designed to generate static scheduling table by static scheduling strategy. This problem belongs to NP-Complete problem, which is firstly solved by the bin-packing algorithm. Secondly, genetic algorithm is used to resolve this problem, but some disadvantages such as early convergence, poor partial searching capability exist in genetic algorithm. Thirdly, this article combines the two algorithms to form the hybrid genetic algorithm to quickly find out the optimal approximate solution. At the end of the article, experiments are made by a set of Automotive Safety System including Adaptive Cruise Control, Electric Power Steering and Traction Control System. Then the effectiveness of the proposed method is verified.