Intelligent elevator control based on adaptive learning and optimisation

[摘要] ENGLISH ABSTRACT: Machine learning techniques have been around for a few decades now andare being established as a pre-dominant feature in most control applications.Elevators create a unique control application where traffic flow is controlledand directed according to certain control philosophies. Machine learning techniquescan be implemented to predict and control every possible traffic flowscenario and deliver the best possible solution. Various techniques will be implementedin the elevator application in an attempt to establish a degree ofartificial intelligence in the decision making process and to be able to haveincreased interaction with the passengers at all times.The primary objective for this thesis is to investigate the potential of machinelearning solutions and the relevancy of such technologies in elevator controlapplications. The aim is to establish how the researchfield of machine learning,specifically neural network science, can be successfully utilised with thegoal of creating an artificial intelligent (AI) controller. The AI controller isto adapt to its existing state and change its control parameters as requiredwithout the intervention of the user.The secondary objective for this thesis is to develop an elevator model that representsevery aspect of the real-world application. The purpose of the modelis to improve the accuracy of existing theoretical and simulated models, bymodulating previously unknown and complex variables and constraints. Theaim is to create a complete and fully functional testing platform for developingnew elevator control philosophies and testing new elevator control mechanisms.To achieve these objectives, the main focus is directed to how waiting time,probability theory and power consumption predictions can be optimally utilisedby means of machine learning solutions. The theoretical background is providedfor these concepts and how each subject can potentially influence thedecision making process. The reason why this approach has been difficult toimplement in the past, is possibly mainly due to the lack of adequate representationfor these concepts in an online environment without the continuousfeedback from an Expert System. As a result of this thesis, the respectiveonline models for each of these concepts were successfully developed in orderto deal with the identified shortcomings.The developed online models for projected waiting times, probability networksand power consumption feedback were then combined to form a new IntelligentElevator Controller (IEC) structure as opposed to the Expert Systemapproach, mostly used in present computer based elevator controllers.

[发布机构] Stellenbosch University