In the bearing only tracking (BOT) problem the observer tries to estimate the position of the target. The observer maneuver is a requirement to guarantee the uniqueness of the estimation. Determining an observer maneuver for BOT is a difficult problem due to its nonlinear nature. This paper proposes a new method that combines the reference BOT scenarios as a priori knowledge with the incoming bearing measurements to obtain an observer maneuver in a generic optimization framework. The use of a priori knowledge makes the observer maneuver design more realistic. The maneuver recommendation process uses the imitation of competing constraints. In this paper, an approach called programming by demonstration is used to recommend a maneuver to the observer. Simulations show that a programming by demonstration framework manages the constraints successfully and generates realistic observer maneuvers.

In the bearing only tracking (BOT) problem the observer tries to estimate the position of the target. The observer maneuver is a requirement to guarantee the uniqueness of the estimation. Determining an observer maneuver for BOT is a difficult problem due to its nonlinear nature. This paper proposes a new method that combines the reference BOT scenarios as a priori knowledge with the incoming bearing measurements to obtain an observer maneuver in a generic optimization framework. The use of a priori knowledge makes the observer maneuver design more realistic. The maneuver recommendation process uses the imitation of competing constraints. In this paper, an approach called programming by demonstration is used to recommend a maneuver to the observer. Simulations show that a programming by demonstration framework manages the constraints successfully and generates realistic observer maneuvers.