An indoor localization method with dual network architecture based on parameter constraints

Abstract

The invention belongs to the technical field of indoor positioning, and in particular relates to an indoor positioning method based on parameter constraints with a dual network architecture. The present invention uses a dual-network architecture based on parameter constraints to extract data features in different fields through different networks, breaking the limitation that a single network architecture can only extract common features, and can fully extract data features in different fields. The linear constraints imposed on the network parameters explicitly model the data distribution drift in the indoor positioning environment, and linearly compensate the distribution drift from the perspective of parameters to minimize domain differences, thereby enabling the model to effectively adapt to complex indoor positioning environments. surroundings. The invention can effectively reduce the difference of data distribution in different fields, so the invention is a method capable of realizing high-precision positioning in complex indoor environments.

Description

technical field [0001] The invention belongs to the technical field of indoor positioning, and in particular relates to an indoor positioning method based on parameter constraints with a dual network architecture. Background technique [0002] With the popularization of smart devices and the rapid development of Internet of Things technology, indoor positioning technology has gained great market opportunities. The growing demand for positioning services based on indoor environments in commercial, medical and military applications has stimulated the rapid development of indoor positioning technologies and systems. Common indoor positioning technologies include infrared, ultrasonic, visible light, UWB, and WiFi, among which infrared, ultrasonic, and visible light positioning requires the deployment of signal transmitters in advance, which requires a lot of manpower and financial resources, so the penetration rate is low; UWB positioning equipment is expensive, usually It is o...

AI Technical Summary

Problems solved by technology

However, the above two methods use the same network structure for feature extraction on data in different fields, which makes the network model only able to mine the common features of data in different fields, which greatly limits the feature extraction ability of the network for data in a certain field.

In addition, since the features extracted by the same network are the common parts of the two domains, the subsequent narrowing of the domain differences will be insufficient, especially when the data distribution of the two domains differs greatly, which will lead to a decrease in the positioning performance of the model.

Based on the above reasons, it is difficult for such methods to achieve accurate positioning in complex indoor positioning environments.

Images