Research On Vehicle Operation And Safety Intelligent Perception Method Based On Multimodal Information Fusion And Adaptive Evolutionary Neural Network

As the most potential means of transportation in the transportation industry,automobile facilitates the human travel mode and promotes the development and exchange of human beings.However,with the increase of vehicle ownership,more traffic congestion and frequent traffic accidents appear.At present,the environment of automobile is becoming more and more complex.Traditional cars rely on driver fatigue to deal with dynamic,changeable and complex environment.The main problem is the lack of the ability to perceive,judge,reason,and make decisions on complex information.In this paper,the research on automobile operation and safety state perception based on multi-modal information fusion and adaptive evolutionary neural network has been carried out based on the research results of human brain cognition by researchers at home and abroad in recent years.In order to realize the intelligent perception of the automobileundefineds operation and the safety state of the occupants,it is necessary to provide the necessary intelligent sensing method to improve the intelligent perception and decision-making ability of the automobile to cope with the complex running environment.Therefore,this paper describes the improvement of vehicle safety performance from the complexity of vehicle traffic environment,the limitations of non-intelligent vehicle in response to the complex traffic environment,and the solutions made by intelligent vehicle in response to these limitations.The cognitive mechanism of the brain is applied to deal with the complex running environment and to solve the problem of complex safety state.The model and description of the intelligent vehicle system operation and safety intelligence perception problem are carried out.Aiming at the defect of single modal information representation ability,high chance and easy to be influenced by the outside world,a multi-modal information fusion method based on evolutionary neural network is proposed,which considers the multi-modal information fusion of human-vehicle system as a whole,and proposes a multi-modal information fusion method based on evolutionary neural network.The intelligent perceptual information processing model of intelligent agent system is constructed.In order to improve the accuracy and speed of automobile perception in complex environment,an evolutionary neural network optimized by adaptive genetic algorithm is proposed for multi-modal information fusion to ensure that the system can effectively utilize complex information in the face of complex information.In order to identify the safety status of the crew,the length-aspect ratio of the driverundefineds eye is obtained by machine vision.The fatigue degree of the driver is judged by the number of blinks for a long time,and the age and gender of the passengers are judged by constructing a Face Net architecture based on the optimization of the deep convolution neural network.Combined with the safety state parameters of the occupants and the self-state parameters of the vehicle driving,an intelligent sensing method of automobile operation and safety is realized through adaptive evolutionary neural network and domain knowledge base.In this paper,the design and development of the intelligent simulation test system for automobile operation and safety state is carried out.Firstly,the driverundefineds fatigue state and the passengerundefineds age and sex status identification can be correctly identified.Combined with the collected data set of vehicle running state parameters,the vehicle operation and safety state perception test is carried out.The self-adaptive evolutionary neural network method is used to identify the safety driving grade of the automobile.The accuracy of the final state perception is 92.48%.Compared with other algorithms,the proposed intelligent vehicle perception method and model are scientific and reasonable.