Autonomous navigation model parameter adaptive optimization method based on edge cloud collaborative framework

The invention relates to an autonomous navigation model parameter adaptive optimization method based on an edge cloud collaborative framework. The method comprises the following steps: S1, constructing a cloud side autonomous navigation model; s2, model parameter issuing: issuing parameters of the cloud side autonomous navigation model to the unmanned aerial vehicle to form a side autonomous navigation model; s3, constructing an online data set; step S4, uploading an online data set; s5, testing the model; s6, outputting an overall error; and S7, updating judgment: setting an error threshold value, comparing an error with the set threshold value, and setting an updating frequency of inputting the online data set into the cloud side autonomous navigation model in the step S5. And the cloud server determines the updating frequency of the training parameters of the cloud side autonomous navigation model of the cloud server based on the error data, so that the calculation overhead of the cloud server is reduced,