Uporaba paralelnega robota pri vstavljanju izdelkov v škatle

PRELOŽNIK, MATJAŽ (author), Munih, Marko (mentor)

V magistrskem delu je predstavljena visoko produkcijska pakirna linija, zgrajena iz osmih postaj, za namen pakiranja, tj. pobiranje polizdelkov s tekočega traku in vstavljanje v škatle. Peta postaja je zaradi kompleksne oblike polizdelkov in kratkih ciklov delovanja izvedena kot robotska celica. Robotski sistem za pobiranje in odlaganje polizdelkov v škatlo temelji na komercialni aplikaciji za pakirne sisteme ABB PickMaster3 in je povezana z robotskim krmilnikom ABB IRC5 v konfiguraciji MultiMove. Krmilnik kontrolira dva delta robota ABB IRB 360 in je razširjen z moduli za nadzor treh tekočih trakov. Vhodna tekoča trakova transportirata dva tipa poljubno razvrščenih polizdelkov do robotskega sistema, na izhodnem tekočem traku pa se nahajajo škatle, v katere urejeno vstavljamo polizdelke. Na podlagi analize tehničnih zahtev za pakirno linijo se je kot edini primerni tip robotskega mehanizma, glede na kratke cikle pobiranja, izkazal delta robot. V delu so zato podrobno podane kinematične lastnosti delta robotov. Druga dva pomembna sklopa robotske celice sta še ustrezno robotsko prijemalo in uporaba strojnega vida za določanje lege izdelkov na tekočem traku. Nastavitve in programiranje robotskega sistema sta bila izvedena v okolju ABB PickMaster3. Aplikacija s pomočjo robotskega krmilnika vodi dva robota v tako imenovanem multi-move independent načinu ter upravlja s tremi tekočimi trakovi hkrati, katere je moč tudi umeriti. Pomemben del aplikacije je industrijski strojni vid za razpoznavanje in določanje pozicije ter orientacije polizdelkov na tekočem traku. Osnovna metoda razpoznavanja je zaznavanje robov, katere na višjem nivoju povežemo v objekte. Za zanesljivo prepoznavanje je ključna dobra kalibracija kamer, ki se izvrši znotraj aplikacije. Parametre sistema smo za dano robotsko celico za doseganje višjega števila ciklov in boljše stabilnosti delovanja na koncu še optimizirali.

• Language: Slovenian
• Keywords: delta robot, direkta kinematika, inverzna kinematika, Jacobijeva matrika, algoritmi računalniškega vida, zaznavanje robov, tekoči trakovi.
• Work type: Master's thesis
• Organization: FE - Faculty of Electrical Engineering
• Year: 2016
• PID: 20.500.12556/RUL-85160
• Publication date in RUL: 14.09.2016

Secondary language

• Language: English
• Title: Use of parallel robot for inserting products into boxes

Abstract

The master thesis presents high-productivity packaging line that collects semi-finished products from the conveyor belt and packs them into boxes. The line consists of eight stations. Due to complex shape of intermediate product and short production cycles is the fifth station designed as a robot cell. Robotic system for picking the objects and sorting in the boxes is based on the commercial application for packaging, ABB PickMaster3, which is connected to controller ABB IRC5. The controller that consists of two delta robots ABB IRB360 is expanded with modules to regulate three conveyor belts. Two input conveyor belts feed the robotic system with randomly positioned and oriented semi-finished products, the third output conveyor belt carries boxes for packaging. Our analysis, considering short picking cycles shows that the only robot configuration satisfying technical requirements for the packaging line is the delta robot. Therefore, detailed kinematic properties of delta robots are presented in the thesis. Two other important components of the robotic cell are the robotic gripper and machine vision for detection of semi-finished products on the feeder belts. Programming and setup of robot cell is implemented in ABB PickMaster3 environment. The application controls both robots in "multi-move independent" mode, and simultaneously controls all three conveyor belts. It also supports calibration of the conveyor belts. Important part of setup is the machine vision, which separates different types of semi-finished products and determines their position and orientation on the conveyor belt. The method of object recognition is based on edge detection, these are being linked into objects. Calibration of cameras is crucial and as such performed within the application. All parameters are also optimized to guarantee higher number of production cycles, highest reliability and efficiency of the packaging line.

• Keywords: Delta robot, direct kinematics, inverse kinematics, Jacobian matrix, computer vision algorithms, edge detection, conveyor belt.