Robot autonomous positioning method based on convolutional neural network fused with IMU and WiFi information

Abstract

The invention discloses a robot autonomous positioning method based on a convolutional neural network fused with IMU and WiFi information. According to the method, a two-channel convolutional neural network is adopted, the method comprises the steps of respectively taking information of WiFi and IMU as inputs of two channels, extracting features through convolution, then automatically obtaining importance degrees of the two channels through an SE module, carrying out weight feature matching, and then obtaining final output through a full connection layer and a softmax function; and taking theserial number corresponding to the reference point as the output of the network, and training the network. In the positioning stage, WiFi and IMU information of a point to be measured is input into atrained network, and the position of a positioning point is estimated by using reference point coordinates corresponding to a serial number output by an output layer and the probability correspondingto the serial number. According to the method, errors caused by RSSI data fluctuation and secondary integration of acceleration in IMU positioning in traditional WiFi positioning can be effectively avoided, and in the positioning stage, the position of the robot can be easily and efficiently obtained.

Description

technical field [0001] The present invention relates to the field of robot positioning, in particular to the integration of an inertial measurement unit (InertialMeasurement Unit, IMU) and WiFi positioning with a two-channel convolutional neural network embedded in an excitation (Squeeze-and-Excitation, SE) module to obtain the position of the robot method study. Background technique [0002] Robot positioning technology is the most basic link to realize autonomous positioning and navigation. It is the position of the robot relative to the global coordinate system and its own posture in the two-dimensional working environment. At present, positioning technology can be divided into two types: absolute positioning and relative positioning: the purpose of absolute positioning is to obtain the position of the positioning target in the global coordinate system, such as WiFi positioning. Relative positioning needs to know the position and orientation of the positioning target at ...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problem that the accuracy of WiFi positioning is easily affected by the environment and the single placement angle of the receiving device, the IMU positioning has cumulative errors, and the positioning accuracy is poor for a long time, the present invention proposes to use a two-channel convolutional neural network (Squeeze-and- Excitation Two channels Convolutional Neural Network, SETCNN) integrates the information of WiFi and IMU to perform robot positioning. The network structure is as attached figure 1 shown

Images