Design And Implementation Of Distributed Guests Recognition System For Automobile 4S Stores

With the development of national economy, people’s demand for cars becomes greater than before. Meanwhile, because of constantly emerging of many kinds of cars’ brands and manufacturers, the more fierce market competition comes out, which requires the Automobile Sales Service shops 4S (4Ss) to do more and more work to satisfy customers’ demands. Customer information, which is generated by the extending car user groups, is the significant guidance to car sales, products launch, and market promotion and so on for 4S shops’ managers. However, for today’s 4S stores, these data are counted by man mostly, so that it can’t be acquired and updated in time and accurately, which lead to lower work efficiency for salesmen and worse car-buying experience for customers. On the other hand, because of the expanding scale of 4S store and geography location distribution of each sub-branch, the resource of each sub-branch is inevitably scattered as well as the staff is difficult to coordinate with each other, then, in the end it is certain that inaccurate statistical data, untimely data synchronize and share will appear, which lead to inaccurate products recommendation to consumers and untimely product promotion to the market.The distributed guests recognition system in the thesis,which is based on 4S stores environment,realizes lots of functions, such as gathering customers’face automatically, recognizing customers from local and neighbor 4S stores, summarizing recognition results, ultimately gathering the incoming users’ information. This thesis is mainly about system framework design of the distributed guests recognition system, and design and implementation of recognition system in each equivalent node. Then, some solutions are proposed to problems which may exist in the detail design of the system. In the system design, P2P structure is adopt to distribute users’recognition tasks to near 4S stores. When recognizing customers, all local computer resources can be divided into management node and computing node, then, management node will distribute recognition tasks from local and neighboring 4S stores to local computing nodes to recognition operation. On the aspect of finding equivalent nodes, P2P can filtrate the unnecessary network communication by the design of discovering algorithm. Management node is in charge of receiving the register of equivalent nodes, registering with equivalent nodes, receiving the register of computing nodes, managing messages and tasks as well as distributing tasks to neighbor nodes and computing nodes. Computing node is responsible for registering with management node, receiving tasks from management node and sending back recognition results. The distributed guests recognition subsystem, the gathering subsystem and the management subsystem communicate under unified user-defined communications protocol. Via the use of this system, the sales staff can easily get customers history information,through these information, the sales man will understand the customer’s specific needs and concerns of the product,then,they can provide targeted service for customers. Marketing department can also get the customers’data easily, and analysis some car models’ attention degree, then carry out more accurate market positioning and product promotion..This thesis describes the research background and significance of the distributed guests recognition system in 4S stores, expounds the domestic and overseas research status in this field firstly. Then analyzes the requests of the distributed system systematically. After that,it elaborates the design of the system structure in detail, including the hardware and software composition and network topology of the system. Later, it divides the recognition system into management and computing nodes, and describes the design and implementation of the system respectively. Finally,this paper introduces system’s test and verification, and summarizes the performance of the system, then analysis the disadvantages during the design and implementation of this system as well as future researches.