Research On Key Technologies Of Vehicle Intelligent Service System In Internet Of Vehicles

With the development of intelligent transportation technology,sensors are gradually integrated into all aspects of travel,and the huge amount of information generated has injected strong vitality into various travel information services and applications.The Internet of Vehicles and sensor technology provide the vehicle with the ability to obtain information,making the vehicle an information carrier,continuously obtaining information from the internal and external environment during the driving process,and feeding it back to the driver and various facilities to support driving navigation,traffic management,etc.However,the traditional service model relies on the demand initiated by users,and the method of"people looking for services" makes the information and service resources ineffective,and also makes the drivers trapped in heavy traffic unable to enjoy the benefits brought by data and information technology.Therefore,how to make full use of this information to proactively provide accurate and personalized transportation services for drivers has become a major challenge for the current intelligent transportation systems.In order to provide drivers with personalized services in the intelligent transportation environment,this research is conducted from four aspects including situation modeling,situational reasoning,situation assessment and personalized customization of services,thus proposing a complete intelligent service system in the Internet of Vehicles environment.The system has important theoretical significance and practical value.The main work and contributions of this paper are as follows:(1)Aiming at the problem of semantic expression of situational contexts and sensor observation capabilities in safe driving situation,an ontology model oriented to entity observation status is constructed.Focusing on several important types of entities in the process of context awareness,concepts and relationships such as people,vehicles,environments,sensors,and states are defined,and semantic relations between sensors,observation objects,and observation results are constructed.By introducing the semantics sensor network ontology,the semantic expression ability of the context model is enhanced for the description of sensor-related concepts and attributes from the aspects of sensor deployment process,observation capability,and observation process.The analysis results of context instances show that the model can effectively support the context expression in the smart service system,laying a solid semantic foundation for subsequent context reasoning,context assessment and personalized service customization.(2)To solove the problems of situational context acquisition and sensor validation,a context reasoning method is established through indepth analysis of the safe driving situation.Starting from the ontologybased context reasoning,the context reasoning part in the system is analyzed and designed from the perspective of sensor observation.By defining the first-order logic semantic reasoning rules,the logic preprocessing of the situation and the ontology reasoning of the situational situation are realized.In addition,in order to check the validity of the sensor observation process,this paper proposes a sensor validation approach based on ontology reasoning.The validity of sensor data is logically expressed by defining corresponding reasoning rules and sensor semantic metadata such as sensor measurement capability and operational constraints.The experimental results show that the reasoning method proposed in this paper can extract the situational context of the driver,vehicle and the environment in the safe driving process,and can also track the validity of the sensors that perform the observation,which provides context support for the subsequent context assessment and service selection.(3)A context assessment model is proposed to address the uncertainty of the observation result caused by the observer(sensor),the observation environment and the observation object during the observation.In the validity evaluation stage,based on the inference results of the validity of each sensor’s measurement attribute,the context validity is calculated according to the weight of each measurement property to the situation observation.The sensor measurement error model is introduced into the evaluation of the context certainty.According to the dividing boundary in the reasoning rules of the situation context,the certainty of the context can be determined.The situation assessment stage determines the situation buffer coefficient based on the dependability of the situation context,thus filtering the situation contexts whose buffer area is lower than the threshold value.Experiments show that the model can adjust the observed status according to the reasoning results within the validity period,which can reduce invalid state changes,improve the efficiency of service invocation control,and reduce the interference of uncertain contexts to the users and the systems.(4)In the aspect of personalized customization of business services,to solve the problems of service composition reusability and muti-context response sequence in the existing approaches,a situation-based service invocation control strategy is proposed.In the context preprocessing stage,for the changing situation context,the criticality of the situation is evaluated based on the context value,the priority and the overall dependability of the situation context.In the service selection stage,the corresponding service combination is acquired according to the changed situation context.The composition strategy and the configuration parameters of each atomic service form an executable service combination.For the problem of multi-context invocation control,a service invocation control algorithm based on the criticality of the context is proposed.By analyzing the criticality of the changed context based on the priority,serverity and dependability of the context,the service composition corresponding to the situation with higher criticality is preferentially executed.By decoupling context from service strategy and service configuration,the reusability of service composition strategy and the flexibility of parameter configuration can be effectively improved,and a feasible technical solution for future context-based service customization can be provided.A hardware prototype was designed and implemented,and the applicability of the inference module on the terminal prototype was evaluated based on two context models.Experiments show that the proposed context model and reasoning method can meet the demands for reasoning performance in mobile scenarios.