Sandesh Thapa - Academia.edu (original) (raw)
Uploads
Papers by Sandesh Thapa
Journal of Intelligent & Robotic Systems, 2019
In this paper, we consider multiple quacopter aerial robots and develop a decentralized adaptive ... more In this paper, we consider multiple quacopter aerial robots and develop a decentralized adaptive controller to cooperatively manipulate a payload. We assume that the mass of the payload is not available to the controller. The developed decentralized adaptive controller employs a consensus algorithm based on connected graphs to ensure that the estimated mass from every agent adds up-to the actual mass of the payload and each agent gets an equal share of the payload's mass. Our controller ensures that all quadcopters asymptotically converge to a constant reference velocity. It also ensures that all of the forces applied to the payload converges to desired set-points. Desired thrusts and attitude angles are computed from the control algorithms and a low-level PD controller is implemented to track the desired commands for each quadcopter. We validate the effectiveness of the controllers in numerical simulations.
Journal of Intelligent & Robotic Systems, 2019
In this paper, we consider multiple quacopter aerial robots and develop a decentralized adaptive ... more In this paper, we consider multiple quacopter aerial robots and develop a decentralized adaptive controller to cooperatively manipulate a payload. We assume that the mass of the payload is not available to the controller. The developed decentralized adaptive controller employs a consensus algorithm based on connected graphs to ensure that the estimated mass from every agent adds up-to the actual mass of the payload and each agent gets an equal share of the payload's mass. Our controller ensures that all quadcopters asymptotically converge to a constant reference velocity. It also ensures that all of the forces applied to the payload converges to desired set-points. Desired thrusts and attitude angles are computed from the control algorithms and a low-level PD controller is implemented to track the desired commands for each quadcopter. We validate the effectiveness of the controllers in numerical simulations.