Unmanned mine car combination positioning optimization method based on probability box

The invention discloses an unmanned mine car combination positioning optimization method based on a probability box. The method comprises the following steps: acquiring original data by using a vehicle-mounted IMU of an unmanned mine car; a probability box P-box is constructed based on a direct modeling method according to the collected original data, features of the original data are extracted and positioned, and feature vector sets in the longitude direction and the latitude direction are constructed by using effective feature vectors; performing feature fusion on the obtained feature vector sets in the longitude direction and the latitude direction according to the dual-channel PCNN, forming a cumulative uncertainty feature matrix, inputting the feature matrix with a label into a normalized exponential function for classification, and checking the precision of a PCNN model based on test set data to obtain a final PCNN model; and according to the obtained final PCNN model, unmanned mine car combination posi