Research On The Technology Of Robotic Book Shelving And Lying-down Book Detection Based On RFID

The management work in the library include inventory and shelving.With the increasing collection,it has been a difficult task to manage the books efficiently and accurately simply relying on manual methods.For the inventory,it is almost an impossible task that the librarians confirm the status of each book only by themselves every day.There are misplaced books in various regions in the library more or less,most of them lie down above the other books which stand on the shelves.It is extremely inefficient to check all the library's books manually just for these small volumes of lying-down books.For the shelves,there are a large number of books returned by readers daily.It is a cumbersome process for librarians to confirm the information of the returned books and the bookshelves they should be.In the process of the shelving,the chaos of the books and a single bookcase layout will often bring trouble to the book shelving,and even lead to the wrong location of the book.How to quickly and accurately return these books back to the original bookshelves so that the follow-up readers can borrow them as soon as possible,has been a traditional problem for the library.In recent years,RFID technology has been popularized in the library.Although there has already been some relatively intelligent RFID equipment to help the librarians manage the books,they still cannot reduce the manual operation,for example,the handheld device.What is more,because of the level of librarians or the complexity of equipment,it is too difficult for librarians to use the equipment.Some equipment are just put in the corner and out of use,this is a really serious waste.In order to solve this problem,this paper presents a library automation management robot based on RFID technology.We have achieved two technologies in this device:book posture detection technology and automatic book shelving technology,to solve the problem of inefficient library management.As a new function of the inventory system,book posture detection technology can help the librarians detect the additional lying-down books above the other books.It recognizes the gestures of the books without consuming manpower.In this paper,we use classification algorithms of machine learning to train the model of books'posture based on the characteristics of RFID tags'signals:read length,read number and change rule of RSSI signal.We use the model to calculate the posture of unknown books,so that we can remind the librarians of the lying-down books' information and correct them.Experiments show that the overall accuracy of the book posture detection algorithm can be up to more than 90%and the average accuracy rate is 80%or more.So we can achieve a relative accurate identification of the attitude of the books.The main advantage of the automatic book shelving system is that the robot can autonomously navigate to the shelves according to route created by the system.It uses the robot technology and RFID technology,stops automatically near the shelves which have books to be put on in the order of the shelves in the route.The whole process doesn't need to manually identify books and shelves,which reduces the burden on librarians and greatly improves the efficiency of the shelving.In addition,the fixed locations of the books bring convenience to readers to find them.Finally,through experimental verification,the positioning accuracy of the system is about 5cm,and it provides the maximum degree of convenience to library staffs on the work of book shelving.