In this paper, we have presented a reliable simulation of algorithm for the recovery of the fiber Bragg grating (FBG) parameters from their reflection spectrum. An accurate method for synthesizing the physical parameters of the fiber Bragg grating has been proposed and demonstrated on the bans of its reflectivity. The method is based on gradient optimization algorithms and can be applied over each type of gratings (uniform, apodized, chirped, phase shifted). We have proposed a new technique, which has overcome some of existing disadvantages and restrictions by using information about the reflection characteristics from grating, which we have examined, and also using the transmission characteristics. As an example, the proposed technique may be successfully applied to synthesize and characterize our experimental grating which is uniformed, apodized and linearly chirped, from FBG measured reflectivity. This method could be applied for distribution strain and temperature sensing applications if only we have some previous information about the
unstrained FBG and also if the possible strain profiles are known. Our algorithm for the synthesis of FBGs from reflectivity has no limitations in parameter space. We don't have to reduce the problem to small number of parameters. In most of practical cases, our technique would fast lead to accurate synthesis results. The proposed method can efficiently lead to optimal solutions and at the same time it takes into account various requirements of the examined grating. This study and simulations are supported by initial laboratory experiments, and this allow us to suppose, that there is a real possibility of complete, fast, and accurate characterization (including phase characterization) of FBG structures from reflectivity measurements.