Dependability Analysis And Optimization Of CBTC Wireless Communication Systems In Urban Rail Transit Systems

In urban rail transit systems,the Communication-Based Train Control(CBTC),which is the widely used signal system,can use high capacity and bi-directional wireless communication system to achieve the effective data transmission.As a safety-critical system,trains in CBTC have high requirements for safety-critical information transmitted by wireless communication system.However,the disturbances in wireless communication system,such as channel fading and wireless security attacks,etc.,can affect the dependability performance of wireless communication,which will affect the authenticity and availability of safety-critical traffic.The decrease of wireless communication authenticity and availability affects the operation efficiency of CBTC systems as well,which will pose a threat to the safe operation of trains.Therefore,optimizing the dependability of wireless communication system,reducing the disturbances,and improving the performance of CBTC systems have become the urgent problem in CBTC systems.The thesis analyzes the requirements of the safety-critical traffic transmitted by wireless communication system,and the necessity to study dependability of wireless communication system.Based on dependability theory and data requirements from users in CBTC systems,including integrity,timeliness,authenticity and availability,the theoretical framework of system dependability is established to describe the integrity,timeliness,security and safety of wireless communication system in urban rail transit systems.After completing the definition of wireless communication system dependability,Ao I(Age of Information)theory is used for the quantitative method of wireless communication system dependability.In order to analyze the effect in wireless communication communication dependability due to environment or human disturbances,overall Ao I,average Ao I and peak Ao I serve as the quantitative indicators of wireless communication communication dependability.Then,after analyzing characteristics of wireless communication in urban rail transit systems,the train-centric CBTC wireless communication system through LTE-T2T(Long Term Evolution-Train to Train)wireless communication technology is introduced based on the modification of LTE-V2X(LTE-Vehicle to Everything).Moreover,to optimize the CBTC wireless communication system dependability,the effective schemes are proposed in this thesis.Machine learning is also used to calculate suitable policies to optimize wireless communication dependability.Finally,based on optimization of wireless communication dependability,a CBTC train control model based NCS(Network Control System)method considering dependability disturbances is established to perform quantified and improved research of relationship between wireless communication communication dependability and CBTC performance.The innovations of this thesis are as follows:1.According to basic feature of dependability,the framework and research status of dependability in engineering computation systems and wireless communication systems are introduced from attributes,threats and means.Based on the distinct features of CBTC wireless communication system and the requirements of safety-critical traffic from users in urban rail transit systems,the theoretical framework of CBTC wireless communication system dependability is defined and analyzed from integrity,timeliness,security and safety.The Ao I theory,including overall Ao I,average Ao I and peak Ao I,is used to quantify CBTC wireless communication system dependability.The effect in wireless communication dependability due to different disturbances are analyzed at last.2.According to basic feature of wireless communication in urban rail transit,a train-centric CBTC wireless communication system through LTE-T2 T wireless communication technology is introduced based on the modification of LTE-V2 X.In order to optimize integrity and timeliness of CBTC wireless communication system dependability,DS-SPS is served as resource allocation scheme in LTE-T2 T scenario,where trains can complete the resource sensing,selection and reservation by themselves instead of track-side e NB.The quantized Ao I is used as quantitative indicators of wireless communication dependability.Machine learning techniques especially Q-learning is further utilized to improve system dependability performance.Simulation results show that the proposed LTE-T2 T based wireless communication system with DS-SPS can achieve improved system dependability and data performance compared with traditional systems.3.In order to solve the Sybil attack threat in CBTC wireless communication system,local security certification scheme and cooperative security check scheme are designed in this thesis to improve the security of CBTC wireless communication system.The local security certification scheme is applied to guarantee the security at the link establishment phase.In cooperative security check scheme,trains can determine whether the received suspicious information is true or not with the help of track-side e NB through three stages of check.Therefore,trains can detect the Sybil attack and recover from the impact from that as soon as possible.Policy-based asynchronous reinforcement learning is utilized to optimize success rate of security check through calculating suitable security check policy and link selection policy.Simulation results show that proposed cooperative security check scheme with policy-based asynchronous reinforcement learning can detect and defense Sybil attack with much higher probability.Thus,CBTC wireless communication system dependability and and data performance can be improved in security side.4.The NCS method is proposed in this thesis to establish CBTC train control model considering dependability disturbances and perform quantified research of relationship between wireless communication system dependability and CBTC performance.In order to optimize the integrated objective function of wireless communication system dependability and CBTC system performance,machine learning method is applied to calculate the suitable communication policy,security check policy and link selection policy.Simulation results show that the velocity vs.distance curve,the number of unplanned braking times and acceleration rate can be optimized through the dependability optimization scheme proposed in this thesis.Therefore,the performance of CBTC system can be improved through our research.