Campus trajectory forecast based on human activity cycle and Markov method

Traditional Markov method used in trajectory prediction fails to capture the property of the moving objects. In this paper, a zoning method was discussed to extract the most popular areas in the campus. We presented a prediction model based on the students´ activity cycle in campus. Markov method was applied in a periodically way to forecast the campus trajectory. Our forecast result was obtained by the weighted integration of different sub-models. Experimental results show that the optimized prediction gives us a satisfying forecast result.