Multi-floor indoor positioning method based on Softmax regression multi-classification recognizer

Abstract

The invention discloses a multi-floor indoor positioning method based on a Softmax regression multi-classification recognizer. The multi-floor indoor positioning method comprises an offline stage andan online stage. The offline stage comprises an offline fingerprint database acquisition stage and an offline multi-floor discrimination classifier construction stage, and the offline fingerprint database acquisition stage establishes an offline RSS fingerprint database by acquiring RSS values of APs sensed by sampling points. In the offline multi-floor discriminant classifier construction stage,a multi-floor discriminant classifier is trained, a cross entropy loss function of the multi-floor discriminant classifier is minimized through a gradient descent algorithm, and an optimal multi-floordiscriminant classifier is constructed by using model parameters corresponding to a minimum cross entropy loss function value. In the online stage, when a to-be-positioned target is positioned, firstly, the probability that the target belongs to each floor is calculated by adopting a multi-floor discrimination classifier, the floor corresponding to the maximum probability is the floor where the to-be-positioned target is located, and then the coordinate position where the to-be-positioned target is located is calculated by utilizing an improved KNN algorithm.

Description

technical field [0001] The invention belongs to the technical field of wireless local area network and indoor positioning, in particular to a multi-floor indoor positioning method based on WiFi and Softmax regression multi-classification identifier. Background technique [0002] The progress of the times has stimulated people's desire for accurate location anytime and anywhere, and in the era of the Internet of Things, this desire has been further intensified. Therefore, the indoor positioning system has penetrated into many aspects of modern life, such as indoor navigation, in-store shopping guide, item tracking and other related activities. In recent years, with the large-scale deployment of wireless communication systems in indoor environments, indoor wireless positioning technology has attracted more and more research interest, and various wireless indoor positioning technologies emerge in an endless stream, such as wireless LAN WiFi, Bluetooth, Sensor networks, ultra-w...

AI Technical Summary

Problems solved by technology

However, with the increase of AP (Access Point) and the number of floors, the computational complexity of this method also increases

Second, the built-in barometer of the smartphone can detect the height of the ground, but not all smartphones have an air pressure sensor, and the movement mode of people in the scene of stairs is very complicated. Different, the accuracy of floor judgment is low

Images