This repository contains the code for "Joint Prototype Learning and Denosing Diffusion Probabilistic Model for Trajetory Imputation", which introduces a novel method for imputing missing points in human trajectory data. The method integrates prototype learning and diffusion models to capture and embed human behavior patterns, making it well-suited for handling incomplete and irregular trajectory data.