Decentralized adaptive flocking control algorithm with avoiding collision and preserving connectivity for crowded UUV swarm with uncertainties and input saturation
This paper investigates a decentralized adaptive flocking control for crowded unmanned underwater vehicle (UUV) swarm in the presence of uncertainties and input saturation. Consider a realistic model of small-size and low-cost UUV with limited communication, it is assumed that only one UUV can acces...
Uloženo v:
| Vydáno v: | Ocean engineering Ročník 237; s. 109545 |
|---|---|
| Hlavní autoři: | , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
Elsevier Ltd
01.10.2021
|
| Témata: | |
| ISSN: | 0029-8018, 1873-5258 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Shrnutí: | This paper investigates a decentralized adaptive flocking control for crowded unmanned underwater vehicle (UUV) swarm in the presence of uncertainties and input saturation. Consider a realistic model of small-size and low-cost UUV with limited communication, it is assumed that only one UUV can access global information (desired path) and each UUV only communicates with its neighbors. To keep all UUVs connected in a crowded swarm, the center of flocking (COF) of each UUV is identified using bio-inspired consensus, thereby a leader-follower flocking controller is proposed to guarantee the collision avoidance and connectivity preservation with bounded fuzzy potential fields at the kinematic level. An adaptive neural networks (NNs) controller is further developed to track the desired path, where an additional control term is incorporated to handle input saturation at the kinetic level. Stability analysis demonstrated that all closed-loop signals are uniformly ultimately bounded. The main contributions of this paper are three-folded. (i) A decentralized algorithm without a prior connectivity assumption on a directed spanning tree is proposed. (ii) Collision avoidance and connectivity preservation can be achieved simultaneously. (iii) An adaptive controller which is robust against model uncertainties and disturbances is developed. Simulation results illustrate the effectiveness of the proposed approach.
(1)Different from the existing leader-following works involved with a directed spanning tree (Cui et al., 2010), this paper introduces a decentralized leader-following flocking control under undirected topology for crowded UUV swarm, where each UUV obtains global information from its neighbors, thereby providing more flexibility and expandability.(2)To keep UUVs connected in a crowded swarm, the COF information is estimated via a novel bio-inspired consensus with the minimum number of neighbors and their optimal spatial distribution, rather than a fixed number of neighbors or distance (Liang et al., 2021a). A bounded fuzzy potential field strategy is subsequently developed to keep the flocking without any potential splits and collisions.(3)An adaptive NN controller is devised for UUV subject to uncertainties and disturbances, in which a saturation compensator is designed to handle the input saturation problem. Furthermore, it is proved that all closed-loop signals are uniformly ultimately bounded with a great convergence in comparison to the existing works. |
|---|---|
| ISSN: | 0029-8018 1873-5258 |
| DOI: | 10.1016/j.oceaneng.2021.109545 |