Recently, a vast amount of research has been conducted on health monitoring of existing structures
such as buildings, bridges and other civil structures. Furthermore, in Japan, natural disasters like typhoons
and earthquakes occur frequently increasing the importance of the damage assessment of the existing
structures. In order to evaluate the damage state of structures, health monitoring technology is quite promising
to provide useful information. However, there are still some research needs in modeling, analysis and
experimental examination before routine applications of health monitoring systems. In this paper, an attempt
is made to develop a damage detection approach system by the learning ability. This learning ability facilitates
a monitoring paradigm without a need for preliminary investigation of the underlying structure and environment.
In other words, it is not necessary to use the precise modeling and analysis methods before conducting
the health monitoring. The proposed system learns the vibration response by using AdaBoost
technique that uses fuzzy-neural networks as a weak learner. By using AdaBoost technique, the network can
respond to various types of external forces and the prediction accuracy increases. The fuzzy reasoning predicts
the next state of structural behavior such as displacement, velocity and acceleration from the current
state of structural behavior and external force. Previously, a health monitoring system that can adapt to the
structural systems and environments through the learning ability was developed with the recognition rate of
over 80% using numerical simulations. However, experimental verification is needed before real life application
of the proposed system. In this paper, results from laboratory experiments are presented to show the effectiveness
of the methodology. It is observed that the proposed system can recognize the change of structural
characteristics and condition states of a large scale steel grid type laboratory structure.