An Improved Weighted K-Nearest Neighbor Positioning Method in Buildings for Power Distribution Room

Accurate indoor positioning is the key technology to realize the “last kilometer” all equipment rapid maintenance of distribution room in buildings. Aiming at the disadvantage that the typical fingerprint positioning method based on single similarity feature is difficult to adapt to the changes of real environment, this paper proposes an entropy weighted k-nearest neighbor (EWKNN) method. Firstly, based on the environment distribution and location characteristics, this paper applies a variety of similarity metrics to optimize the nearest neighbor of the target node. Secondly, in view of a large number of data samples in fingerprint database, for the shortcoming of weight process that relies too much expert experience, the paper uses entropy weight method to adjust the weight proportion of each index adaptively. Finally, the optimal predicted position of the target node is obtained. Meantime, Simulation results show that the proposed EWKNN method improves errors by 9% and 11% compared to other methods, respectively. It validates the correctness of the theoretical analysis and the effectiveness of the proposed higher precision indoor positioning strategy.