Dynamic Coverage Control in a Time-Varying Environment Using Bayesian Prediction
针对未知时变密度函数的多智能体覆盖控制问题，提出基于贝叶斯预测的密度估计方法和离散控制方案，使智能体逼近最优部署，并分析了均方稳定性。
This paper investigates the dynamic coverage control problem for a group of agents with unknown density function. A cost function, depending on a certain metric and the density function, is defined to describe the performance of coverage network. Since the optimal deployment of agents is closely depending on the density function, we employ the Bayesian prediction approaches to estimate the density function. Moreover, a novel coverage-control-customized algorithm is proposed to acquire the Bayesian parameters. The merits of this Bayesian-based spatial estimation algorithm are the consideration of measurement noise and the capability of dealing time-varying density function. However, the estimated density function from Bayesian framework follows normal distribution, which leads the cost function to a stochastic process. To deal with this type of cost function, a discrete control scheme is proposed to steer the agents approaching to a near-optimal deployment. The mean-square stability of the proposed coverage system is further analyzed. Finally, numerical simulations are provided to verify the effectiveness of the proposed approaches.