为解决高速公路基础设施运行监测和健康管养人工成本高、效率低的问题，面向高速公路巡检应用场景提出了多无人机协同路径规划方法。以路径长度、能量消耗和路径平滑所构成的综合巡检代价最小为优化目标，构建了包含任务分配、无人机性能和巡检飞行等约束条件的三维路径规划模型。通过建立空间模型模拟高速公路巡检环境，设计仿真实验应用蚁群算法、遗传算法和模拟退火算法分别求解多无人机协同路径规划问题。结果表明：随着巡查点规模增大，遗传算法在场景三（120个巡查点）中相较蚁群算法和模拟退火算法巡检代价分别降低0.52%和0.39%，展现出更优的大规模任务处理能力；在任务分配方面，遗传算法实现了更均衡的续航利用率，使得3架无人机续航里程利用率依次为79.00%、59.10%和46.50%；在综合性能上，遗传算法优化效率和能耗水平相对更优，其规划路径总长度和最长无人机路径长度分别为55.4 km和23.7 km，最大电池消耗为79%；尽管收敛速度较慢，但遗传算法的全局搜索能力强，其最优解质量与迭代次数比值最低，仅为0.12，且不易陷入局部最优。因此，遗传算法更适用于高速公路巡检多无人机协同路径规划任务，可提升算法优化效率、降低飞行能耗水平、改善任务分配均衡性，为高速公路多无人机协同巡检提供基础技术支撑。

To address the challenges of high operational costs and low efficiency in operation monitoring and health management of expressway infrastructure, this paper proposes a multi-UAV collaborative path planning method tailored to expressway inspection scenarios. A three-dimensional path planning model is formulated with the objective of minimizing the comprehensive inspection cost of path length, energy consumption, and path smoothness, while incorporating critical constraints such as task allocation, UAV performance, and inspection flight. A spatial simulation environment is constructed to replicate realistic expressway inspection conditions. Comparative simulation experiments are conducted using three metaheuristic algorithms—ant colony optimization, genetic algorithm, and simulated annealing—to solve the multi-UAV collaborative path planning problem. The results confirm that as the number of inspection points increases, in Scenario 3 (120 inspection points), the inspection cost of the genetic algorithm decreases by 0.52% and 0.39% respectively compared with the ant colony algorithm and the simulated annealing algorithm. The genetic algorithm performs better in handling large-scale and complex tasks. In terms of task allocation, the genetic algorithm achieves a more balanced range utilization rate, with the endurance utilization rates of the three UAVs being 79.00%, 59.10%, and 46.50% respectively. In terms of overall performance, the optimization efficiency and energy consumption level of the genetic algorithm are relatively better. The total planned path length and the longest UAV path length are 55.4 km and 23.7 km respectively, with a maximum battery consumption being 79%. Although its convergence speed is slow, the genetic algorithm has strong global search capabilities, with the lowest ratio of optimal solution quality to iteration times, only 0.12, and is not easily trapped in local optima. Therefore, the genetic algorithm is more applicable for multi-UAV collaborative path planning in expressway inspection, which can improve optimization efficiency, reduce flight energy consumption, and improve task allocation balance, providing fundamental technical support for multi-UAV collaborative inspection on expressway.