Unmanned aerial vehicles (UAV) are increasingly used in more challenging and demanding surveillance and reconnaissance (S&R) missions. This leads to higher system complexity and requires more sophisticated solutions. In addition, in machine-based aerial S&R with imaging sensors, changing environmental and operating conditions as well as varying parameter settings often result in reduced detection performance. This imposes high demands on sensor capability and perception algorithms on board UAVs. To optimize detection and tracking performance, it is necessary to closely coordinate the movement of the sensor-bearing platform, sensor alignment and actual sensor data acquisition, taking into account platform, target and environment characteristics. We present a concept for automated sensor-specific trajectory generation for UAVs, under consideration of numerous constraints. By adjusting the UAVs trajectory, sensor-related geometric dependencies (such as target range, bearing and elevation) can be adapted specifically to facilitate data acquisition in a way that subsequent perception algorithms exhibit improved detection and tracking performance. This approach includes the capability of tracking multiple spatially separated targets observed by a single UAV.
展开▼
