Show simple item record

dc.rights.licenseRestricted to current Rensselaer faculty, staff and students. Access inquiries may be directed to the Rensselaer Libraries.
dc.contributorMishra, Sandipan
dc.contributorGandhi, Farhan
dc.contributorJulius, Anak Agung
dc.contributorHicken, Jason
dc.contributor.authorHu, Botao
dc.date.accessioned2021-11-03T09:01:04Z
dc.date.available2021-11-03T09:01:04Z
dc.date.created2018-07-27T15:11:31Z
dc.date.issued2018-05
dc.identifier.urihttps://hdl.handle.net/20.500.13015/2216
dc.descriptionMay 2018
dc.descriptionSchool of Engineering
dc.description.abstractUnmanned Aerial Vehicles (UAVs) are finding widespread popularity in many applications and as a result, new research challenges and solutions are being discovered at a remarkable rate. One of the big barriers that prevent UAVs from wider deployment is realizing full autonomous flight. Autonomous systems typically consist of guidance, navigation and control subsystems. The guidance system plans and generates trajectories that guide a UAV to complete its mission. The navigation system senses the environment and estimates the relative or absolute state of the UAV in a local or global map. The control system then generates control commands based on the state estimation results and the reference trajectory. In this dissertation, we specifically focus on trajectory generation (as part of the guidance subsystem) for the time-optimal landing of a UAV onto a moving platform.
dc.description.abstractFinally, we propose and validate a mechanism for generating a time-optimal trajectory for maneuvering a quadrotor through a constrained and dynamically changing the environment. Though the algorithm developed has broader applicability and is generic, maneuvering a quadrotor passing through a moving window (translating and tilting) is used as an example scenario. The position and Euler angles of the window pose time-varying yet partially defined intermediate state constraints. To address the issue of evolving and time-varying constraints, the algorithm divides a trajectory into sub-trajectories and uses the intermediate state constraints as boundary state constraints for each sub-trajectory. Experimental validation and simulation benchmark results illustrate the effectiveness of the proposed trajectory generation algorithm.
dc.description.abstractNext, we extend this idea to a general time-optimal trajectory generation algorithm for landing a 6-DOF quadrotor model onto a translationally moving and rotating platform. This algorithm exploits the differential flatness of the quadrotor dynamics model and formulates a nonlinear programming problem, which is then solved to obtain the time-optimal landing trajectory. The advantages of the proposed algorithm over state-of-the-art solution techniques for time-optimal trajectory design include computational efficiency and the ability to incorporate dynamics and state constraints (such as collision avoidance from an obstacle) into the optimization problem. The proposed algorithm is applied to a 3-DOF quadrotor model for generating the landing trajectory and a feedback controller is designed to deal with disturbance and modeling uncertainty. Experimental results for landing a quadrotor onto a heaving and rolling platform show that the proposed algorithm is able to achieve time-optimal landing and is capable of also avoiding obstacles during the landing maneuver.
dc.description.abstractlanding controller demonstrates that the proposed control architecture is able to achieve time-optimal landing using only onboard sensor measurements.
dc.description.abstractWe first present a general control architecture for time-optimal UAV landing onto a moving platform. The control architecture consists of a motion estimation module, a trajectory generation module and a tracking control module. The motion estimation module estimates the motion states of the platform and the quadrotor simultaneously. With these estimates and enforced motion constraints (such as maximum acceleration, maximum velocity, etc.), the trajectory generation module generates a time-optimal reference landing trajectory. The tracking control module then uses an adaptive robust controller (ARC) in the inner loop to track the reference trajectory. We validate the effectiveness of the control architecture by demonstrating the time-optimal landing of a quadrotor onto a heaving platform. Experimental results and comparison with a state-of-the-art
dc.language.isoENG
dc.publisherRensselaer Polytechnic Institute, Troy, NY
dc.relation.ispartofRensselaer Theses and Dissertations Online Collection
dc.subjectMechanical engineering
dc.titleTime-optimal trajectory generation for autonomous landing
dc.typeElectronic thesis
dc.typeThesis
dc.digitool.pid179055
dc.digitool.pid179056
dc.digitool.pid179057
dc.rights.holderThis electronic version is a licensed copy owned by Rensselaer Polytechnic Institute, Troy, NY. Copyright of original work retained by author.
dc.description.degreePhD
dc.relation.departmentDept. of Mechanical, Aerospace, and Nuclear Engineering


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record