Design Of System Of Vehicle Maintenance Management Based On Julius Speech Recognition Technology

In 2019,the total number of vehicles in China will reach 260 million,and the number of vehicles in the world will break through a new high year after year,and the vehicle maintenance market will increase day by day.However,in the maintenance process of traditional vehicle maintenance industry,there are many problems,such as paper-based maintenance sheet pollution,easy to miss,complex maintenance process,low maintenance efficiency and so on.In order to solve the above problems,a vehicle maintenance system is designed based on speech recognition technology,which realizes paperless and electronic maintenance process,reduces maintenance cost,simplifies maintenance process and improves maintenance efficiency.The vehicle maintenance system is based on C / s client server model architecture.Based on WPF framework,Julius speech recognition engine,Microsoft speech synthesis API,springboot development framework and Tomcat server technology,the client and server of the vehicle maintenance system are completed.The vehicle maintenance process control is completed in the client and the vehicle maintenance result storage is completed in the server.Based on the realization of the vehicle maintenance system,the speech recognition performance of the system is optimized.Finally,the problem of digital recognition difficulty is solved by using DNN acoustic training model.The recognition results of different maintenance types are isolated by using dictionary segmentation,and the recognition rate of single maintenance task is improved by adding voice feedback.The speech recognition performance of the system reaches a higher standard It improves the availability of the system.In the process of training DNN model,we find a method to reduce the number training set by using key phonemes.Through the combination of key phonemes,137 numbers are used to cover the 10000 numbers from zero to 99,which greatly reduces the size of training set when training DNN model.The test results show that the final vehicle maintenance system has higher speech recognition rate and better use experience,which can help simplify vehicle maintenance process,improve maintenance efficiency and promote paperless vehicle maintenance.