Personalized Service Strategy And Algorithm In Ambient Intelligence

Ambient intelligence (AmI) is an interconnected system which is user-centric, intelligent and personalized and also a new type of interactive mode established between the people and the environment. In nature, AmI is the integration of ubiquitous computing, ubiquitous communication and user interface technologies; it provides people with an electronic environment with intelligent characteristics.Users can enjoy the ubiquitous, transparent and personalized information services in AmI and thereby the center of their attention will return to the task itself.Personalized service is to take the initiative to meet user's personalized needs starting from the difference of their preferences, knowledge structure and behavioral pattern. Personalized service strategy is not only an important component, but also one of study objectives and core values of AmI theory.It is of great significance for further enriching and perfecting AmI theory to study and propose a set of personalized service strategy which can be consistent with the characteristics of AmI and take full advantage of the related technologies. In this paper, we consider the application scenarios of exhibition and exposition and then propose a set of overall design scheme of personalized service strategy in AmI, which is based on Bluetooth multi-antenna system. We abstract the strategy to two parts (user division and service recommendation) and use Fuzzy C-Means and item-based collaborative filtering algorithm Slope One as the basic algorithm of these two mechanisms respectively. We also introduce the feedback and update mechanism based on user's residence time and the data fusion mechanism based on improved D-S evidence theory into our personalized service strategy. On the one hand, Slope One algorithm is suitable for off-line operation and doesn't have steep demand on the amount of new user's preference information, so it can be used to solve the problems of scalability and user cold start; on the other hand, we can get a better solution to the problems of sparsity and item cold start by integrating the content information of items into the Slope One algorithm. Furthermore, we adopt Bluetooth short-range wireless communication technology and Field Programmable Gate Array (FPGA) technology to design a Bluetooth multi-antenna system, which is composed of a microprocessor and a number of Bluetooth modules, and we will take it as the Service Node of our personalized service strategy in AmI. As a result of the work carried out in this paper, AmI theory has been further enriched and improved.