Interior Planning and Design Analysis Considering the Improvement of PDR Positioning Technology

Abstract

To solve the problem of insufficient indoor positioning accuracy, a motion recognition and positioning method based on improved gait detection is proposed. In this method, the data is collected by an acceleration sensor, and the plane step estimation and vertical distance estimation algorithms are used to identify and analyze the features of different motion states. A one-dimensional convolutional neural network is used to improve the accuracy of step size estimation in the process of going up and down stairs. Comparative experimental results show that the total positioning errors of Pedestrian Step Estimation and Vertical Estimation algorithms are 0.605 m and 0.367 m, respectively. The total errors of the traditional Route Planning Algorithm and the Non-dominated Sorting Genetic Algorithm-iii algorithm are 3.071 m and 2.316 m, respectively. The experimental results show that the 1D-CNN algorithm has obvious advantages in the case of non-synchronous length, and the positioning errors in the X, Y, and Z axes are 0.298 m, 0.187 m, and 0.103 m, respectively, indicating that the proposed method significantly improves the accuracy of position estimation in the indoor environment.