室内时变流场中多机器人源定位方法实验与数值验证

为了快速和准确定位室内时变流场中的污染源，提出了一种综合的多机器人源定位方法（URPSO），并综合利用机器人实验和数值实验来验证该方法的有效性。首先，在风扇左右周期性摆动的实验环境中，利用3台机器人开展了15组独立性实验，其中14组实验成功定位源，成功率达到93.3%，说明URPSO方法在室内时变流场中具有强鲁棒性。其次，针对真实实验环境，利用仿真方法开展了机器人源定位实验，结果表明:仿真方法和实验方法获得的成功率一致，且2种方法的平均定位步数和方差接近，说明利用仿真方法开展源定位研究是可行的。最后，利用仿真方法在3种典型室内时变流场环境（混合通风案例MV、置换通风案例DV和自然通风案例NV）中，分别开展了100组独立性实验，对应的成功率分别为100%、92%和81%，说明URPSO方法在不同时变流场环境中均具有较高成功率。

EXPERIMENTAL AND NUMERICAL VALIDATION ON A MULTI-ROBOT SOURCE LOCALIZATION METHOD FOR DYNAMIC INDOOR ENVIRONMENT

To quickly and accurately locate contaminant sources in dynamic indoor environments, this study presented a multi-robot source localization method (URPSO) by integrating concentration and airflow information and validated the method by combining robot experiments with numerical simulations. The experimental study was first conducted by using three mobile robots to locate an ethanol source in a typical dynamic indoor environment with a fan swinging periodically from left to right. A total of 14 out of 15 experiments were successful, with a success rate of 93.3%, indicating that the method had strong robustness. Next, according to the experimental environment, numerical robot experiments for locating the ethanol source were conducted. The results showed that the success rate and the average number of steps from numerical experiments were consistent with those from robot experiments, indicating the feasibility of using numerical simulations. Finally, 100 numerical experiments were further conducted by using URPSO method to locate an ethanol source in three typical dynamic indoor environments (mixing ventilation case MV, displacement ventilation case DV and natural ventilation case NV), respectively. For MV, DV and NV, the success rates were 100%, 92% and 81% respectively, indicating that URPSO method had high success rates in different dynamic indoor environments.