Design And Implementation Of Voice Assistant Central Control System

With the rapid development of Internet technology,voice assistants have widely entered our lives.To be able to use voice to manipulate a wide variety of devices,many manufacturers want to incorporate voice assistants into these devices.In order to improve the reusability of voice assistants,this paper proposes to design a universal voice assistant system for voice operations.Various intelligent devices can achieve the corresponding operations by calling the system's interface to obtain the sentence analysis results.Due to the increase in the number and source of requests,in order to improve the concurrency capability,the voice assistant adopts a distributed design and optimizes the request forwarding rate through the central control system to ensure the integrity of the data during transmission.Central control system uses the play framework to build the framework and uses SBT tools to manage the project.Compared to traditional J2 EE projects,this tool can naturally integrate scala and java projects.The functional programming provided by scala naturally supports high concurrency scenarios.At the same time,the concurrency framework that comes with scala can make up for the lack of lightweight Java concurrent tools.The system forwards and processes requests through a unified interface to ensure the integrity and security of data transmission.Before the request data is returned to the user,the system uses Redis to construct a cache between sentence keywords and execution vertical regions.When the next request for a similar sentence comes,the query cache is used to obtain the execution vertical region,which reduces the access of irrelevant vertical regions and improves the access speed of similar statements.At the same time,in order to ensure that the request can be processed in a timely manner,the system uses Nginx to distribute the request according to the vertical domain.During the distribution,the server's load balancing is achieved through ip hash.Finally,for some special sentences,the meaning of the sentence cannot be understood correctly through semantic understanding,and the system will directly judge by direct speech.This design can not only speed up the reaction rate of the system,but also make the return result of the system more accurate.The central control system controls the logical forwarding and scheduling of eachpart of the voice assistant through a detailed interface,thereby ensuring the integrity of the data during transmission.The direct speech module assists in semantic understanding,which makes the accuracy of sentence processing higher.Finally,the statement execution efficiency is improved through the load balancing service and cache mechanism.