Font Size: a A A

Schedule Optimization Of A Robot Production Line And Design Of Experiment Platform

Posted on:2018-09-17Degree:MasterType:Thesis
Country:ChinaCandidate:S L ZhangFull Text:PDF
GTID:2348330512979860Subject:Control engineering
Abstract/Summary:PDF Full Text Request
As a major equipment for digitized workshop and intelligent factory,robotic automation production line plays a crucial role in intelligent manufacturing.Robot handling production line is one of the typical types,which is widely used in package classification,component extraction and cargo handling.So the robot automatic production line scheduling optimization problem,which can make system operation cost lowest in long-term case,has important significance for the actual production.Robot handling production line consists of a multi-degree of freedom robot for picking and placing,one or more convey belt for conveying work piece and industrial visual configuration system used for work piece positioning,recognition,measurement,etc.The convey transports work pieces to the robot work range in some arrive rate.The robot decides to wait,pick or place work pieces with some strategy,according to work piece position information and buffer spare capacity.Our optimization objective is to find a optimal strategy,which can make the system operation cost lowest in long-term case.This thesis setups a Discrete-Time Markov Decision Process(DTMDP)optimization model,considering the average time which robot picks a work piece as system takt time,and considering the robot buffer capacity and work piece position information as state.Next,we use the theory optimization method and simulation optimization method to solve the DTMDP model,through policy iteration algorithm and the Q-learning algorithm based on Simulated Annealing.And furthermore,we set the optimal parameters for system according to the simulation results.In order to simulate and research the robot production line in actual production better,this thesis designs a kind of robot production line experiment platform,including robot system,conveying system,vision system,air supply system,HCI centralized control system and intelligent environment support system.The experiment platform accomplishes the capture of work piece information,picking and placing of the work piece and auto circulation of work piece box.The experiment platform supplies experiment measures to simulation and research of production work mode and control method for laboratory.
Keywords/Search Tags:Robot handling production line, Discrete-Time Markov Decision Process, SA-Q-learning, Production line experiment platform
PDF Full Text Request
Related items