| 摘要 |
A geographical region having an object to be tracked is divided into subregions, where adjacent disposed subregions partial overlap one another and the overlapping portions of adjacently disposed subregions share common states. A transition matrix is used to represent the terrain in a subregion and each subregion can have N geographical states and four (4) additional transition states. A regional tracker is used to estimate state sequences for each new observation of the object. Tracking continues until the process is complete or until the probability that the vehicle is in another new subregion occurs. Tracking performed between adjacently disposed subregions is based, at least in part by, the direction represented by the transition state. Tracking into a new subregion from an adjacently disposed old subregion begins precisely where tracking for the old subregion ended using the states that are common in the overlapping portions of the adjacently disposed subregions. |
