Land-mark placement for reliable localization of automatic guided vehicle in warehouse environment

The reliable and effective localization system is the basis of Automatic Guided Vehicle (AGV) to complete given tasks automatically in warehouse environment. However, there are no obvious features that can be used for localization of AGV to be extracted in warehouse environment and it dose make it difficult to realize the localization of AGV. So in this paper, we concentrate on the problem of optimal landmarks placement in warehouse so as to improve the reliability of localization. Firstly, we take the practical warehouse environment into consideration and transform the problem of landmarks placement into an optimization problem which aims at maximizing the difference degree between each basic unit of localization. Then Genetic Algorithm (GA) is used to solve the optimization problem. Then we match the observed landmarks with the already known ones stored in the map and the Triangulation method is used to estimate the position of AGV after the matching has been done. Finally, experiments in a real warehouse environment validate the effectiveness and reliability of our method.