Research And Implementation Of TTCAN Network Based On Improved Genetic Algorithm

CAN bus compared to the traditional field bus in the reliability,real-time and flexibility aspect has great advantages,and CAN bus is widely used in the automotive industry,aerospace,shipbuilding,industrial control,sensors and other fields.While applying CAN protocol,some event trigger mechanism flaws has been detected.The defects will lead to CAN bus performance degreadation in the systems which have real-time,stability requirements.In order to solve the CAN protocol flaws,the time-triggered TTCAN protocol is proposed in this dissertation and more and more people attach importance to it.TTCAN adopts the method of dividing the time window to guarantee the real-time quality of the periodic message,and set up the windows of the arbitration window to balance the transmission of the Non-periodic message.Similarly,the TTCAN protocol is also have flawes that the system matrix optimization problem is too complex and can not be applied to the dynamic network.Based on the TTCAN protocol and genetic algorithm,this dissertation improves the TTCAN system matrix and system deployment update mode,completes the algorithm design and experimental verification.Specific content is as follows:1.An improved genetic algorithm to optimize the TTCAN system matrix is proposed.This algorithm can effectively improved the real-time performance,reliability and bandwidth utilization of TTCAN.Based on the new encoding method which is combination of matrix coding and real coding,the algorithm of construction initial system matrix is designed,which solves the irrational problem of completely randomly generating individual.At the same time,the steps of crossover and mutation are changed.2.A dynamic update mechanism of the TTCAN system matrix is proposed.Solved the problem that the traditional TTCAN system matrix can not be applied to the dynamic network due to the static offline pre-generation.This mechanism allows it to automatically regenerate new system matrix and deploy to the system when the system network changes.3.Design and implement the experimental verification platform.Use this platform to verify the improved genetic algorithm and dynamic update mechanism.The experimental verification platform is a multi-node TTCAN network,which consists of an embedded development platform(host computer)and a node using the STM32F105 as the MCU.the QT master control program,and CAN driver based on embedded Linux and STM32 environment are designed.4.The performance analysis of the improved genetic algorithm is carried out on the experimental verification platform,And the performance of dynamically updating TTCAN system matrix is tested.Finally the hardware system's basic functions,real-time and error rate preformance is tested.The experimental results shows that the proposed algorithm improves the bandwidth utilization rate by 4.3% for the international standard PSA news set in the automotive industry,increases the sending chance of 73 random messages per second within the arbitration window,the reliability of the network is improved.And the dynamic update mechanism can ensure that the bandwidth utilization is always maintained at a high level when a network node is added or deleted,can meet the requirements of applications in the dynamic network.