In order to eliminate effects of non-line-of-sight（NLOS）error on positioning precision of ultra-wideband（UWB）indoor positioning system

an UWB indoor positioning algorithm based on convolutional neural network（CNN）is proposed.By using UWB system to collect the indoor positioning data and according to the error characteristics of signal propagation in the NLOS environment

a CNN model is established and the data is input into the network for training

so as to reduce the influence of NLOS error.Combined with the inertial navigation system（INS）

the extended Kalman filter（EKF）is used to fuse the position estimation of UWB/INS system to further improve the accuracy of positioning.The experimental results prove that the proposed algorithm can effectively reduce the influence of NLOS error

and the location-solving trajectory is closer to the real trajectory

which can meet the requirements of centimeter（cm）-level positioning.