UAV Waypoint Opportunistic Navigation in GNSS-Denied Environments
UAV Waypoint Opportunistic Navigation in GNSS-Denied Environments
复制标题
DOI：
10.1109/taes.2021.3103140
复制
发表时间：
2021-08
影响因子：
4.4
通讯作者：
Z. Kassas;Yanhao Yang;Joe J. Khalife;Joshua Morales
中科院分区：
文献类型：
--
作者：
Z. Kassas;Yanhao Yang;Joe J. Khalife;Joshua Morales
考虑了在全球导航卫星系统（GNSS）拒绝环境下，无人机（UAV）导航到达期望航路点的可证明保证问题。假设UAV具有未知的初始状态（位置、速度和时间），并且假设环境具有多个具有未知状态（位置和时间）的地面机会信号（SOP）发射器和一个状态已知的锚SOP。UAV对所有SOP进行伪距测量，以估计其自身的状态以及未知SOP的状态。航路点导航问题被公式化为贪婪（即，一步前瞻）多目标运动规划（MOMP）策略，保证无人机以用户指定的置信度到达航路点的用户指定距离内。MOMP策略平衡了两个目标：（i）导航到航路点;（ii）减少UAV的位置估计不确定性。它表明，在这样的环境中，制定一个所谓的“天真”的方式直接前往航路点的航路点导航问题将导致无法到达航路点。这是由于环境的可估计性差。相比之下，MOMP策略保证（在概率意义上）到达航路点。蒙特卡罗模拟结果表明，MOMP策略实现了预期的目标，95%的成功率相比，36%的成功率与天真的方法。实验结果的无人机导航到一个航路点在一个细胞SOP环境中，MOMP策略成功地到达航路点，而天真的策略未能做到这一点。
Navigation of an unmanned aerial vehicle (UAV) to reach a desired waypoint with provable guarantees in global navigation satellite system (GNSS)-denied environments is considered. The UAV is assumed to have an unknown initial state (position, velocity, and time) and the environment is assumed to possess multiple terrestrial signals of opportunity (SOPs) transmitters with unknown states (position and time) and one anchor SOP whose states are known. The UAV makes pseudorange measurements to all SOPs to estimate its own states simultaneously with the states of the unknown SOPs. The waypoint navigation problem is formulated as a greedy (i.e., one-step look-ahead) multiobjective motion planning (MOMP) strategy, which guarantees that the UAV gets to within a user-specified distance of the waypoint with a user-specified confidence. The MOMP strategy balances two objectives: i) navigating to the waypoint; and (ii) reducing UAV’s position estimate uncertainty. It is demonstrated that in such an environment, formulating the waypoint navigation problem in a so-called “naive” fashion by heading directly to the waypoint would result in failing to reach the waypoint. This is due to poor estimability of the environment. In contrast, the MOMP strategy guarantees (in a probabilistic sense) reaching the waypoint. Monte Carlo simulation results are presented showing that the MOMP strategy achieves the desired objective with 95% success rate compared to a 36% success rate with the naive approach. Experimental results are presented for a UAV navigating to a waypoint in a cellular SOP environment, where the MOMP strategy successfully reaches the waypoint, while the naive strategy fails to do so.