A Drowning Features Recognition System Based On AIS

The computing method inspired by bio can be divided into artificial neural networks, evolutionary computation, artificial immune system (AIS), swarm intelligence and etc. The AIS is a relatively young research field but particularly active in recent years, and has become an important branch of computer science. The AIS is a computational or imformation system based on the mechanism or characteristic of biological immune system to solve engineering problems. It originated from biologists want to expand the immunology theory to the engineering field, scientists try to solve some problems of computer security, fault diagnosis and industrial control and other areas of theoretical research and practical application by simulating biological immune system mechanism or phenomena. The application and study of the AIS complement one another and pursue common development, the research of AIS can be divided into three direction:theory research, system modeling and algorithm research, this paper focus on the study of AIS algorithm.With the popularization of application and the development of research, its defects and limitations are also increasingly prominent, the researchers also according to the application demand raise various improved algorithm. This paper based on the study of the classic clonal selection algorithm and the dynamic clonal selection algorithm found they have poor real-time issue, and proposes a real-time clone selection algorithm to make up the flaw of the existing artificial immune algorithm, and be used and verified in the embedded platform.The main research contents of this paper can be divided into the theory research of AIS algorithm and its application in the embedded platform, the main work is as follows:In the aspect of theory research, put forward the real-time clonal selection algorithm based on the research of the existing AIS model and clonal selection algorithm. The main contents are as follows:①Detector training and detector detecting is no longer a completely independent process, the process of detector training be split into two parts:basic feature study and specific feature study, and embed the later part into the detection stage;②The antibody library own different lengths of antibodies for meeting various antigen’s length, and variable length can also be used as an antibody variation during maturate;③During the process of detection, the algorithem dynamically adjust the dear degree computing priority of antibodies in the antibody library according to various antigen, so as to reduce the test time.In the application part, clarify the drowning feature recognition system functional requirements. On the base of passed the feasibility analysis, completed the system framework design and the system software and hardware modules division and design, apply the real-time clonal selection algorithm in the embedded system, and verify its accuracy, real-time and reliability. The main work is as follows:①Complete the system function and framework design and modules divide according to the actual demand. The Hardware part can be divided into signal collection and pre-processing module, data pre-processing module, drowning features recognition module and monitor module.②Complete the software and hardware modules’design.③Realize the real-time clonal selection algorithm on the platform.④Verify its accuracy, real-time and reliability.Finally, summarize the main content of the thesis, list the advantages and disadvantages of the system, give other solutions about drowning feature recongnition and put forward expectations about the application of artificial immune system and embedded system combined.