Active Elastic Anticipation+ Model: Swarm flocking through relative velocity and acceleration estimation in 3D

Date

2025-4-10

Author

Keskin, Ersin

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

92
views

0
downloads

Cite This

Natural phenomena such as bird flocking and fish schooling demonstrate how simple local interactions can give rise to organized and efficient collective behavior. These biological insights have inspired swarm robotic systems, where agents coordinate based on the positions and orientations of nearby peers. Beyond reactive behavior, anticipation, which involves predicting the future states of neighboring agents, enables smoother and more coordinated motion. Building on this concept, the Active Elastic Anticipation model was introduced to improve swarm coordination through velocity-based prediction. However, as it is defined only in two-dimensional space, it shows limited performance in complex and dynamic environments. This study proposes an enhanced model, Active Elastic Anticipation+, with three key improvements. First, the anticipation mechanism is extended to incorporate both velocity and acceleration. Second, the model is generalized to three-dimensional environments to handle complex obstacle geometries. Third, a decentralized Extended Kalman Filter is integrated, allowing each agent to estimate its own velocity and acceleration using only local position measurements. These enhancements improve swarm responsiveness, coordination, and robustness. The proposed model was validated through two-dimensional and three-dimensional simulations and real-world experiments using CrazyFlie 2.1 UAVs. In simulations, the model maintained cohesive formations in narrow passages and showed a fast recovery from disturbances. In physical experiments, it successfully preserved swarm structure under sensing constraints, with decentralized estimation enabling accurate and timely state updates. In conclusion, the proposed model advances swarm robotics by enabling more realistic and adaptive behavior. Its improvements support robust operation in real-world scenarios such as search and rescue, exploration, and environmental monitoring.

Subject Keywords

3D Flocking, Anticipation in Collective Motion, Kalman Filter, Swarm Robotics

URI

https://hdl.handle.net/11511/114622

Collections

Graduate School of Natural and Applied Sciences, Thesis