Prediction-based target tracking is a very important method for target location and tracking in wireless sensor networks

however

a common difficulty in traditional prediction algorithm is that the target location prediction is not accurate in real environment

due to dynamic change of target motion mode

etc. To overcome these shortage

propose a grey Markov model-based target tracking（ GMMTT） algorithm. Introduce Markov model with turbulence characteristics into segmentation grey prediction

which can not only achieve good precision of target localization

but also adapt to change of target motion mode. After getting predicted position of target

current tracking nodes propagate tracking information to the predicted area where the target will arrive in the next moment

and wake up nodes in this area in advance

so that effective tracking of target can be achieved with as few nodes as possible. Experimental result shows that GMMTT algorithm has superior performance in tracking probability and tracking precision.