The localization of a radio frequency (RF) emitter is relevant in many military and civilian applications. The recent decade has seen a rapid progress in the development of small and mobile unmanned aerial vehicles (UAVs), which offer a way to perform emitter localization autonomously. The path a UAV travels influences the localization significantly, making path planning an important part of a mobile emitter localization system. The topic of this thesis is path planning for a UAV that uses bearing measurements to localize a stationary emitter. Using a directional antenna, the direction towards the target can be determined by the UAV rotating around its own vertical axis. During this rotation the UAV is required to remain at the same position, which induces a trade-off between movement and measurement that influences the optimal trajectories. This thesis derives a novel path planning algorithm for localizing an emitter with a UAV. It improves the current state of the art by providing a localization with defined accuracy in a shorter amount of time compared to other algorithms in simulations. The algorithm uses the policy rollout principle to perform a nonmyopic planning and to incorporate the uncertainty of the estimation process into its decision. The concept of an action selection algorithm for policy rollout is introduced, which allows the use of existing optimization algorithms to effectively search the action space. Multiple action selection algorithms are compared to optimize the speed of the path planning algorithm. Similarly, to reduce computational demand, an adaptive grid-based localizer has been developed. To evaluate the algorithm an experimental system has been built and the algorithm was tested on this system. Based on initial experiments, the path planning algorithm has been modified, including a minimal distance to the emitter and an outlier detection step. The resulting algorithm shows promising results in experimental flights.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Sensor Path Planning for Emitter Localization


    Beteiligte:

    Erscheinungsdatum :

    28.02.2023


    Anmerkungen:

    Doctoral thesis, UCL (University College London).


    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch


    Klassifikation :

    DDC:    629




    Non-myopic Sensor Path Planning for Emitter Localization with a UAV

    Hoffmann, F / Schily, H / Krestel, M et al. | BASE | 2023

    Freier Zugriff


    Single-sensor RF emitter localization based on multipath exploitation

    O'connor, Alan / Setlur, Pawan / Devroye, Natasha | IEEE | 2015


    Calibration Signal-Assisted Emitter Localization Under Sensor Position Uncertainty

    You, Ming-Yi / Pu, Wenqiang / Zhou, Rui et al. | IEEE | 2025