Research On Context Classifier-based Environmental Awareness Of The EACS System

EACS system is a composite service system which allows service adjustment during execution. The system can monitor the environment, and analyze the monitoring results. When a service quanlity fail to satisfy the users’needs to the whole composite service quantity, appropriate operation can be done to guarantee the whole composite service’s quanlity, such as substitution. For this EACS system, the foundation is the environmental awareness, and the destination of environmental awareness is to abtain the condition of each service in system by analyzing and recognizing the environmental monitoring results. Also, these process described above is the main issue we are concentrating in this thesis.In the analysis of the environmental awareness demand to the EACS system, this thesis proposes an environmental awareness framework which is based on context classifier and choosing HMM as classification and identification model of context classifier. Besides, a context construction method based on key change is proposed which can reduce the HMM matching times of classifier and improve the efficiency.In this thesis, there is a further discussion of the usage of HMM, the structure and construction method of the vector quantizer for HMM data preprocessing, and select the suitable code number of vector quantizer by an experiment; also, this thesis proposes a context construction method based on key change and a series of algorithms like key change analysis, key change creation, training sequence selection. At last, another experiment is given to test the whole method, and therefore, proved the applicability of the whole environmental awareness process.