A UAV Target Prediction and Tracking Method Based on KCF and Kalman Filter Hybrid Algorithm

Wei, Huanxia

doi:10.1109/iccece54139.2022.9712830

Cited by 9 publications

(2 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tuning gains of vertical and yaw controls, i.e., k V h and k Ψ, made it possible to reduce that distortion. Based on Figures 11,12,15,16,20 and 21, it can be stated that values of these gains equal to about 0.5 eliminate or reduce the overregulation significantly. Observing the impact of V max on flight trajectories in all cases of reference trajectory shape, it can be concluded that the maximum relative speed does not have to be greater than the speed of the tracked point.…”

Section: Discussionmentioning

confidence: 97%

“…The present work presents a method that applies an artificial-potential-field approach and modified flight control for quadcopters to achieve the tracking of a freely moving point with unknown dynamics. Thus, in such conditions, MPC (model predictive control) [ 15 ] or Kalman filter [ 16 ] based methods cannot be applied, because the estimation or prediction of the target’s movement within the time horizon is not possible, or the model of the target is not similar to the UAV in contrast to [ 17 ], where sliding mode control is applied. The proposed method can also be used in applications where a virtual moving point is controlled by a ground station to achieve complex shapes of trajectories with the use of any local positioning system.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Artificial Potential Field Based Trajectory Tracking for Quadcopter UAV Moving Targets

Kownacki

2024

Sensors

View full text Add to dashboard Cite

The trajectory or moving-target tracking feature is desirable, because it can be used in various applications where the usefulness of UAVs is already proven. Tracking moving targets can also be applied in scenarios of cooperation between mobile ground-based and flying robots, where mobile ground-based robots could play the role of mobile landing pads. This article presents a novel proposition of an approach to position-tracking problems utilizing artificial potential fields (APF) for quadcopter UAVs, which, in contrast to well-known APF-based path planning methods, is a dynamic problem and must be carried out online while keeping the tracking error as low as possible. Also, a new flight control is proposed, which uses roll, pitch, and yaw angle control based on the velocity vector. This method not only allows the UAV to track a point where the potential function reaches its minimum but also enables the alignment of the course and velocity to the direction and speed given by the velocity vector from the APF. Simulation results present the possibilities of applying the APF method to holonomic UAVs such as quadcopters and show that such UAVs controlled on the basis of an APF behave as non-holonomic UAVs during 90° turns. This allows them and the onboard camera to be oriented toward the tracked target. In simulations, the AR Drone 2.0 model of the Parrot quadcopter is used, which will make it possible to easily verify the method in real flights in future research.

show abstract

Section: Discussionmentioning

confidence: 97%