Trajectory Tracking Based on BP Neural Network and Extended Kalman Filter
In indoor positioning,ultra-wideband(UWB)technology,as a new positioning technology,is widely used in various working scenes.As the environment is complicated and varied,UWB communication signals are easily blocked,which will interfere with the positioning of the tracking target and the judgment of the motion tracking.In or-der to obtain the motion track of the target,in addition to considering whether the collected data is obtained under the condition of signal interference,the system noise filtering should be carried out on the data,and then the movement trajectory of the target can be obtained.In this paper,the problem of discriminating whether the signal is disturbed was transformed into the binary classification problem under machine learning.The collected signals were classified by establishing the BP neural network model,and the data without interference were filtered out.The calculation results show that the training accuracy of the BP neural network is as high as 96.43%and the testing accuracy is 99%.On this basis,in order to get the trajectory diagram of the moving target in practical application scenarios,the extended Kalman filter algorithm was used to filter the system noise,and finally,the trajectory of the tracking target was ob-tained.
NETL
NSTL
万方数据
维普
