A non-ranging positioning method for wireless sensor networks based on stacked autoencoders

Abstract

The present invention relates to a wireless sensor network non-ranging positioning method based on a stacked autoencoder, including: an anchor node broadcasts its own anchor node information to nodes in the network, and the node collects all anchor node information; the anchor node collects the anchor node information The node information constructs the hop count vector and the distance vector and sends them to the sink node; the sink node uses the two received vectors as training data to train the stack autoencoder to obtain the distance estimation model, and sends the model to the unknown node in the network; the unknown node The collected anchor node information is input into the model, and the output of the model is the distance estimation vector from the unknown node to the anchor node; the unknown node constructs the objective function according to the distance estimation vector, and uses the solution corresponding to the minimum value of the objective function as its own coordinates. According to the characteristics of the actual environment of the network, the present invention improves the practicability of the non-ranging positioning method in the deployed network by learning the data collected by the anchor nodes in the network, and realizes precise positioning under different networks.

Description

technical field [0001] The invention relates to the communication field, in particular to a non-distance-measuring positioning method for a wireless sensor network based on a stacked autoencoder. Background technique [0002] Node localization in wireless sensor networks is to estimate the position of nodes whose positions are unknown in the network under the premise that there are several anchor nodes (Anchornodes) with known positions in the network. At present, positioning algorithms in WSNs can be divided into two categories, which are range-based and range-free. The ranging-based positioning algorithm needs to install additional hardware devices on the sensor nodes to collect ranging-related physical parameters, such as angle of arrival (AoA), time of arrival (ToA) and received signal strength (RSS), etc., to measure the transceiver the distance between. This type of method can usually obtain high positioning accuracy, but the installation and use of additional hardwa...

AI Technical Summary

Problems solved by technology

[0004] In the actual deployment of large-scale sensor networks, sensor nodes are usually deployed by low-altitude sowing of aircraft, which will lead to uneven distribution of nodes in the network, and the environment to be monitored usually has obstacles, lakes, etc. Areas where nodes cannot be deployed, thus creating "coverage holes" in the network, such as figure 2 As shown, the emergence of "coverage holes" will lead to the detour of the shortest routing path between nodes. In practice, radio signals are affected by obstacles, fading and multipath effects. In the case of the same transmission distance, the received power of the signal is not all the same, as image 3 As shown, radio irregularity (Radio Irregularity) will also have a negative impact on the positioning scheme that uses the number of hops between nodes for distance estimation

[0005] The above various external influences that are not conducive to positioning are collectively referred to as anisotropic factors, and the network with these factors is called an anisotropic network, and the existing positioning methods are affected by the anisotropic factors in the network. The shortest distance between anchor nodes and the actual distance often have a large difference, so the location information of unknown nodes obtained often has a large error, which leads to inaccurate estimation of the location information of unknown nodes and reduces the positioning accuracy of wireless sensor networks.

The existence of anisotropic factors in wireless sensor networks will have a greater impact on the performance of positioning algorithms, and the anisotropic factors vary greatly in networks of different deployment environments, making the performance of the designed positioning scheme fluctuate more in different networks. It is difficult to achieve good positioning performance in all environments of the network, and the coordinate calculation methods used by most current positioning algorithms have inherent defects, which further limit the positioning accuracy of the algorithm

Images