Hava manipülatörü takımıyla otonom meyve toplama

Manipulation is the ultimate capability for autonomous micro unmanned aerial vehicles (MAVs), which would enable a substantial number of novel use-cases. Precision agriculture is such a domain with plenty of practical problems that could utilize aerial manipulation, which is faster with respect to ground manipulation. Apple harvesting is the most prominent use case with ever-growing percentages in the overall apple production costs due to increasing imbalance between labor supply and demand. Moreover, contemporary industrial UGVs equipped with arms still strive to compete against human labor financially, which is not the case for relatively cheaper aerial manipulators. Nevertheless, the dynamics of the aerial manipulator is significantly more convoluted when compared to plain quadrotors and thus demand a more sophisticated control pipeline. In this thesis, a brief dynamics analysis on the airframe with a 3 DOF manipulator is conducted and subsequently an appropriate autopilot software with an adequately stable control pipeline -PX4- is selected. Thereafter RRT path planning algorithm family, followed by its multi-robot variant sRRT, which uses subdimensional expansion framework, is introduced. Then the inverse kinematics of 3 DOF arm is introduced, which is used alongside with the local 1 planner in fruit picking. Consecutively a supervised apple detection model based on Mask R-CNN and Data Distillation with real-time capable inference is developed. At the end, multi-robot apple picking scenario is realized and the integration of these submodules is verified in simulation.