基于贝叶斯网络控制算法的智能电梯探讨

Intelligent Elevators Based on Bayesian Network Control Algorithm

基于物联网学习算法的智能电梯控制系统在现实生活中对于提高多层建筑的可持续性和便利性方面发挥着重要作用,随着物联网及大数据的智能时代到来,传统电梯控制算法对应随机人流带来的不确定性已略显乏力。与传统的电梯控制方法不同,贝叶斯网络控制算法利用了图像采集系统和预处理系统为其提供的乘客群体大小和等待时间信息样本,然后利用这些信息样本进行概率决策模型的贝叶斯推理、更新变量参数,利用变量消除技术降低了计算边际概率和条件概率的复杂度。其目的是构建一个决策引擎,能够控制电梯的行动,以提高不同楼层、不同时间段和不同需求乘客的满意度。研究表明,根据不同的场景,对该模型算法进行了敏感性分析和评价研究,在94%的测试场景中,整个算法被证明达到了相应的预期期望值。在总结前者研究经验及成果的同时,提出了一种基于贝叶斯网络控制算法的智能电梯工作的基本建模方法,为工程师提供了一种开放式思路应用于实际工作中。

Intelligent elevator control system based on Internet of Things learning algorithm plays an important role in improving the sustainability and convenience of multi-storey buildings in real life.With the arrival of the intelligent era of Internet of Things and Big data,traditional elevator control algorithm has been slightly weak in dealing with the uncertainty caused by random people.Different from the traditional elevator control method,the Bayesian network control algorithm uses the image acquisition and preprocessing system to provide sample group size and waiting time of passengers’information.Then the information samples probability model of Bayesian inference is used to update the variable parameter.The variable elimination technology is used to reduce the complexity of calculating the marginal probability and conditional probability.The aim is to build a decision engine that is capable of controlling the elevator's actions to improve the satisfaction of passengers on different floors,different periods and different needs.Some studies have shown that sensitivity analysis and evaluation research have been carried out on the model algorithm according to different scenarios.In 94%of the test scenarios,the whole algorithm has been proved to meet the corresponding expected expectations.In this paper,the experience and results of the former research are summarized,and a basic modeling method of intelligent elevator based on Bayesian network control algorithm is proposed,which provides an open idea for engineers to apply in practical work.