Hoang Duc N M, Tran Duc M, Tran Thanh-Sang, Pham Hoang-Anh
Faculty of Computer Science and Engineering, Ho Chi Minh City University of Technology (HCMUT), Ho Chi Minh, Vietnam.
Vietnam National University of Ho Chi Minh City (VNU-HCM), Ho Chi Minh, Vietnam.
PeerJ Comput Sci. 2021 Feb 16;7:e388. doi: 10.7717/peerj-cs.388. eCollection 2021.
Cooperative navigation for fleets of robots conventionally adopts algorithms based on Reynolds's flocking rules, which usually use a weighted sum of vectors for calculating the velocity from behavioral velocity vectors with corresponding fixed weights. Although optimal values of the weighting coefficients giving good performance can be found through many experiments for each particular scenario, the overall performance could not be guaranteed due to unexpected conditions not covered in experiments. This paper proposes a novel control scheme for a swarm of Unmanned Aerial Vehicles (UAVs) that also employs the original Reynolds rules but adopts an adaptive weight allocation mechanism based on the current context than being fixed at the beginning. The simulation results show that our proposed scheme has better performance than the conventional Reynolds-based ones in terms of the flock compactness and the reduction in the number of crashed swarm members due to collisions. The analytical results of behavioral rules' impact also validate the proposed weighting mechanism's effectiveness leading to improved performance.
传统上，机器人集群的协作导航采用基于雷诺兹群聚规则的算法，该算法通常使用向量的加权和，根据具有相应固定权重的行为速度向量来计算速度。尽管通过针对每个特定场景的大量实验可以找到能给出良好性能的加权系数的最优值，但由于实验中未涵盖的意外情况，整体性能无法得到保证。本文提出了一种针对无人机群的新型控制方案，该方案同样采用原始的雷诺兹规则，但基于当前环境采用自适应权重分配机制，而不是在一开始就固定权重。仿真结果表明，我们提出的方案在群聚紧凑性以及减少因碰撞导致的集群成员坠毁数量方面，比传统的基于雷诺兹的方案具有更好的性能。行为规则影响的分析结果也验证了所提出的加权机制的有效性，从而带来了性能的提升。
