Tracking-Aided Target Classification Using Multi-Hypothesis Sequential Test

This paper deals with target classification by using both feature data and kinematic measurements. The problem is tackled by multi-hypothesis sequential testing with embedded target tracking. We implement Armitage's sequential probability ratio tests (SPRT) for non-maneuvering and maneuvering targets. Two track fusion architectures, including centralized fusion and distributed fusion, are used to handle the embedded tracking problem. The benefit of the kinematic measurements to classification is analyzed and improvement is shown analytically for a special case. Numerical results are provided to demonstrate the performance of our algorithm.