Quantum particle filter: a multiple mode method for low delay abrupt pedestrian motion tracking

摘要

Pedestrian tracking is a key building block in many emerging technologies, and there are many applications that require a very accurate and reliable pedestrian tracking. Two problems, namely the abrupt changes in motion and the delay in the tracking process are addressed in the reported work. The approach seeks to simulate the uncertainty of the position of electrons around the nucleus by propagating particles in areas where the person is more likely to be, depending on the person's possible motions and taking into account various sources of uncertainty. A novel feature of the presented approach is that it can track abrupt motions accurately with a low delay. The robustness of this approach for abrupt motion situations is demonstrated.