Design And Implementation Of A Smart Book Shelving System Based On RFID

Nowadays,more and more libraries begin to adopt the RFID technology.RFID has many features,which make RFID not only a simple alternative to barcode and magnetic stripe,but also making the library able to provide more advanced services.In the traditional library,the books on the bookshelves are strictly sorted by the call numbers,but the position of each book is not fixed.However,in the RFID Smart Library,each book has a fixed position,so people can get the position of each book by searching from the OPAC system,which provides convenience for the readers who want to find a specific book,but also brings great work load for the librarians.In order to solve this problem,we developed an intelligent book shelving system based on RFID technology,which can greatly improve the efficiency of book shelving work,and reduce the likelihood of mistakes.The main hardware device for the intelligent book shelving system is an RFID intelligent trolley,consisting of an RFID reader with several RFID antennas,a computer host,a screen,and an accumulator.The system can be divided into four function modules,which are book loading and scanning,path generating,navigation,and book shelving alerting.The librarian put the books on the trolley,and then use the RFID antenna to scan all the books on the trolley.According to the RFID tags attached on each book,the system can get the information of each book from the database.Knowing the position of each book,we can use route selecting algorithm to find a path,and displayed it on the screen.When the librarian is walking with the trolley,the RFID reader would read the anchor tags attached on bookshelves,and then calculate the current position of the trolley and display it on the screen,which provides navigation for librarians.When the trolley reached a specific position,the system will pop up a prompt box,showing which books should be put on the shelves here.This study focuses on the route selecting and trolley locating.Route selecting is a classic NP problem,unable to find a solution of polynomial time complexity.This paper presents a cluster-based routing algorithm.The algorithm is compared with backtracking algorithm,greedy algorithm and Ant Colony algorithm in terms of time consuming and route length.We found that the algorithm can find a relatively short route within an acceptable period of time and is more applicable than other algorithms above.For the locating of smart trolley,we learn the relation between max reading distance and RFID reader power,the relation between the RSSI value and reading distance by real-place experiment.Then we can choose an appropriate power for locating algorithm.First,we propose the Single Tag Locating algorithm,in which the power is set relatively small,so we can reduce the influence of multipath effect.The algorithm is very stable,and can meet the basic requirements of the application.But the precision of Single Tag Locating algorithm is only 1m.The improved algorithm is based on K-NN algorithm.Based on the training data acquired from former experiments,we can get a group of RSSI tuples,which respectively represents the RSSI values of four nearest reference tags.By comparing locating result of multiple powers,we find that the locating error is minimum when the RFID reader power is set to 25-26dbm.