移动传感器目标跟踪的连续时间估计感知最优控制
Continuous-Time Estimation-Aware Optimal Control of Mobile Sensors for Target Tracking
- University of Minnesota（明尼苏达大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出一种估计感知最优控制方法，用于移动传感器目标跟踪，考虑非线性动力学和协方差，确保连续时间约束满足，并在稀疏离散化下有效。
AI中文摘要：
目标跟踪是利用（通常由）移动传感器收集的测量值来估计目标系统状态的问题。移动传感器的轨迹优化不仅需要考虑动力学和操作约束，还需要考虑估计目标系统状态时的不确定性。我们提出了一种用于目标跟踪的估计感知最优控制方法，该方法考虑了移动传感器和目标系统的非线性动力学、扩展卡尔曼滤波中的非线性协方差动力学以及非线性测量模型。此外，它确保移动传感器的约束在连续时间内（在数值精度范围内）得到满足。在数值模拟中，所提出的方法即使在稀疏时间离散化下也能实现估计感知的目标跟踪，同时满足连续时间内的移动传感器约束。
英文摘要：
Target tracking is the problem of estimating the state of a target system using measurements collected (oftentimes) by mobile sensors. Trajectory optimization for mobile sensors must account not only for dynamical and operational constraints, but also for uncertainty in estimating the target system's state. We propose an estimation-aware optimal control method for target tracking that accounts for the nonlinear dynamics of both the mobile sensor and target system, the nonlinear covariance dynamics in extended Kalman filtering, and nonlinear measurement models. Furthermore, it ensures constraints on the mobile sensor are satisfied in continuous time (within numerical precision). In numerical simulations, the proposed method achieves estimation-aware target tracking while satisfying the mobile-sensor constraints in continuous time, even under sparse time discretization.
