The Design And Implementation Of Vehicle Assisted-driving System Integrated With Speech Recognition

In recent years,with the rapid development of information technology and computer technology,higher requirements have been made for the convenience and safety of automobiles.However,the existing vehicle-assisted driving system does not comprehensively deal with the detection of speech recognition,lane lines,and road markings.Therefore,in view of this research status,this paper designs a vehicle-assisted driving system that integrates speech recognition,focusing on Drivers' voice signals during driving and recognition of lane lines and traffic markings on driving roads were studied to improve the safety of vehicle-mounted driving assistance systems.Firstly,through a detailed analysis of demand for all types of drivers,through their repeated research to determine the integration of voice-based automotive assistance system research objectives.And through the demand analysis results,combined with the specific features of speech recognition and target detection tasks,the overall design of the system functions.Secondly,we carefully analyze the voice information commonly used by drivers and establish the corresponding datasets as the test data.We propose the Markov selection model based on hidden Markov model HMM.At the same time,considering the influence of environmental factors on speech recognition technology,Speech Recognition Using Driver 's Lips Information.Thirdly,a combined model of linear and clothoid is established.The lane detection is achieved by extracting positive and negative edge points of lane line,voting of line segments based on Hough transform and matching of edge line and lane line.Based on extended Kalman filter and lane departure model,dynamic tracking of lane line and deviation warning function are realized.Based on the idea of learning,the road traffic marking recognition is realized,so that a complete vehicle-assisted driving system is achieved.Finally,the relevant experiments are carried out on the speech recognition and vehicle-assisted driving related methods in this paper,and the verification is carried out from the aspects of accuracy,robustness and real-time performance.