Research on the applications of the genetic algorithm to sensor registration
Ling Wu, Faxing Lu, Peifa Jia · 2007
Sensor registration is a basis for well-organized sensor network, and a precondition for data fusion. In cases of constant registration errors, batch processing methods are always applied, where the registration is actually viewd as an optimization problem. Such methods are fast convergent, but sometimes they are not flexible in different cases and the optimal techniques used in batch processing methods may provide the suboptimum as solution. What’s more, when dealing with a large number of sensors, the batch processing methods may come across numeric problems. To address the registration problem in some practical cases, the evolutionary algorithm based method can be explored. A method based on genertic algorithm, as well as the least squares method, are developed for sensor registration in different simulation scenarios and compared. Simulation results are analyzed to make clear the advantages and disadvantages of the methods.