基于图像的电梯群控系统优化调度的研究

OPTIMAL SCHEDULING OF ELEVATOR GROUP CONTROL SYSTEM BASED ON IMAGE

电梯群控系统作为现代建筑交通系统重要组成部分,以满足用户的呼叫需求为目标,进行电梯集群的调度分配。然而传统电梯群控调度算法未考虑到乘客群体的差异性,平等地为所有乘客分配电梯轿厢,从而导致需要更多空间的特殊乘客需要更长的等待时间。将图像识别技术应用到电梯群控系统中,通过SSD模型估算出电梯门厅中等待乘客的人数及其物品。通过使用从监视各层电梯门厅收集到的检测结果,对电梯群控系统调度进行优化。仿真实验结果表明,相比于传统群控调度算法,该方法可以大幅减少特殊乘客平均候梯时间,减少重复呼梯的情况,进而降低电梯集群的能耗。

As an important part of modern building traffic system,elevator group control system is used to dispatch and allocate elevator group to meet the call demand of users.However,in the traditional elevator group control scheduling algorithm,the differences of passenger groups are not considered,and elevator cars are allocated equally to all passengers,which leads to the special passengers who need more space need longer waiting time.In this paper,image recognition technology was applied to the elevator group control system,and the number of waiting passengers and their goods in the elevator lobby were estimated by SSD model.The elevator group control system scheduling was optimized by using the detection results collected from the elevator lobby of each floor.The simulation results show that compared with the traditional group control scheduling algorithm,this method can significantly reduce the average waiting time of special passengers,reduce the repeated calls,and reduce the energy consumption of elevator cluster.