Abstract: COVID-19 is highly sensitive and highly pathogenic. How to avoid interpersonal contact during distribution of epidemic prevention materials is a major problem for epidemic prevention and control. Considering that the distribution robot is small, fast and adaptable to the environment, this paper establishes a centralized control multi robot collaborative distribution mode with the shortest distribution time as the main goal and the shortest distribution path as the secondary goal. In the global path planning stage, this paper improves the A* algorithm to solve its shortcomings such as many inflection points and long redundant paths. In the local path planning stage, it improves the dynamic window method （DWA） to enable it to choose a better path. Finally, the two algorithms are combined to enable the robot to travel on the preferred path and achieve the effect of autonomous obstacle avoidance during the travel. Finally, the model and algorithm are validated by taking the distribution of medical protective materials in Henan People’s Hospital under the COVID-19 epidemic situation as an example. The results show that the robustness, path length, distribution time and path smoothness of the fusion algorithm have been optimized to varying degrees compared with the traditional algorithm, and this algorithm can provide a reference for the non-contact epidemic prevention material distribution mode of the robot in the hospital.

Key words: “contactless” distribution, epidemic prevention materials, robot, path optimization

摘要： 新型冠状病毒传感性强，致病率高，如何避免防疫物资在配送时人际接触是疫情防控的一大难题。考虑配送机器人具有小快灵且环境适应性强等特点，以配送时间最短为主目标，以配送路径最短为次目标建立集中式控制多种类机器人协同配送工作模式。在全局路径规划阶段，改进A*算法以解决其存在拐点多、冗余路径长等缺点，在局部路径规划阶段改进动态窗口法（DWA）使其能够选择更优路径，最终将两种算法融合使机器人能够达到择优路径行驶并在行驶中达到自主避障的效果。以新冠病毒感染疫情（COVID-19）下河南省人民医院医用防护物资配送为例，对模型与算法进行验证。结果表明，融合算法与传统算法相比，鲁棒性、路径长度、配送用时和路径平滑度都得到了不同程度的优化，该算法可为医院内部机器人无接触式防疫物资配送模式提供参考。

关键词: “无接触”配送, 防疫物资, 机器人, 路径优化