Multisensor Poisson Multi-Bernoulli Filtering with Uncertain Sensor States

In a typical multitarget tracking (MTT) scenario, the sensor state is either assumed known, or tracking is performed based on the sensor’s (relative) coordinate frame. This assumption becomes violated when the MTT sensor, such as a vehicular radar, is mounted on a vehicle, and the target state should be represented in a global (absolute) coordinate frame. Then it is important to consider the uncertain sensor location for MTT. Furthermore, in a multisensor scenario, where multiple sensors observe a common set of targets, state information from one sensor can be utilized to improve the state of another sensor. In this paper, we present a Poisson multi-Bernoulli MTT filter, which models the uncertain sensor state. The multisensor case is addressed in an asynchronous way, where measurements are incorporated sequentially based on the arrival of new sensor measurements. In doing so, targets observed from a well localized sensor reduce the state uncertainty at another poorly localized sensor, provided that a common non-empty subset of features is observed. The proposed MTT filter has low computational demands due to its parametric implementation. Numerical results demonstrate the performance benefits of modeling the uncertain sensor state in feature tracking as well as the reduction of sensor state uncertainty in a multisensor scenario compared to a per sensor Kalman filter. Scalability results display the linear increase of computation time with number of sensors or features present.