A sensor management protocol for tracking with diverse sensors

We develop a sensor management protocol that a vehicle may use to track other vehicles and objects in its neighborhood using a sensor-suite, an individual element of which gives either periodic cluttered updates or a periodic uncluttered updates. We outline how to combine the joint probabilistic data association filter with stochastic Kalman filters for the state estimation process. However, since such an algorithm is sensitive to modeling errors, we discuss the use of data-driven optimization algorithms to increase robustness. Finally, we discuss extensions of previously developed covariance control sensor management algorithms.