We study a generalized motion planning problem involving multiple autonomous robots navigating in a $d$-dimensional Euclidean space in the presence of a set of obstacles whose positions are unknown a priori. Each robot is required to visit sequentially a prescribed set of target states, with the number of targets varying between robots. This heterogeneous setting generalizes the framework considered in the prior works on sequential parametrized topological complexity by Farber and the second author of this article. To determine the topological complexity of our problem, we formulate it mathematically by constructing an appropriate fibration. Our main contribution is the determination of this invariant in the generalized setting, which captures the minimal algorithmic instability required for designing collision-free motion planning algorithms under parameter-dependent constraints. We provide a detailed analysis for both odd and even-dimensional ambient spaces, including the essential cohomological computations and explicit constructions of corresponding motion planning algorithms.
翻译:我们研究了一个广义运动规划问题,涉及多个自主机器人在$d$维欧几里得空间中导航,环境中存在一组位置先验未知的障碍物。每个机器人需要按顺序访问一组预设的目标状态,且不同机器人的目标数量各不相同。这种异构设置推广了Farber与本文第二作者先前关于序列参数化拓扑复杂度的研究框架。为确定该问题的拓扑复杂度,我们通过构造适当的纤维化对其进行数学建模。我们的主要贡献在于广义设定下该拓扑不变量(即参数化拓扑复杂度)的确定,该不变量刻画了在参数相关约束下设计无碰撞运动规划算法所需的最小算法不稳定性。我们对奇维与偶维环境空间均进行了详细分析,包括必要的上同调计算以及相应运动规划算法的显式构造。