ROS-based control of GLUON desktop robot

V diplomski nalogi smo razvili preprosto aplikacijo poberi in odloži z namiznim robotom GLUON. Zasnovali smo preprosto paralelno dvoprstno prijemalo in ga natisnili s 3D tiskalnikom. Za aktuator prijemala smo uporabili elektromotor N20 s kodirnikom in reduktorjem, za njegovo vodenje pa razvojno ploščo STM32 Nucleo L476RG. Za krmiljenje aktuatorja prijemala smo jo nadgradili z obstoječo razširitveno ploščo X-NUCLEO-IHM04A1. Izdelali smo lastno razširitveno ploščo, namenjeno napajanju Nuclea preko omrežja robota in povezavi v komunikacijsko CAN omrežje le-tega. Sestavljeno prijemalo smo kot samostojno enoto integrirali na robota GLUON. CPE Nuclea smo nastavili tako, da je omogočala CAN komunikacijo. Osnovali smo način komunikacije preko CAN omrežja s CANopen protokolom in ustvarili namenska sporočila za vodenje prijemala. Zasnovali in implementirali smo PD regulator za pozicijsko vodenje prstov prijemala. Napisali smo Python knjižnico, ki preko UDP povezave omogoča vodenje prijemala iz zunanjega računalnika. Postavili smo ROS okolje za premikanje robota in ga, s pomočjo generatorja trajektorij Pilz Industrial Motion Planner ter demonstracijskih datotek proizvajalca, naučili premikov. S pomočjo omenjenega generatorja trajektorij smo ustvarili Python knjižnico za premikanje robota in v njej implementirali sistem za učenje točk s pomočjo prehajanja med stanjem izklopljenih aktuatorjev in pozicijskim vodenjem robota. Postavili smo okolje aplikacije in s sistemom za učenje točk robota naučili, kako izvesti aplikacijo ter jo izvedli.The thesis addresses the development of the pick and place application with the GLUON desktop robot. We designed a simple parallel two-wrist gripper and printed it with a 3D printer. We used an N20 electric motor with encoder and reducer for the gripper actuator, and for its control, we used an STM32 Nucleo L476RG development board. We upgraded the gripper actuator with an existing X-NUCLEO-IHM04A1 expansion board to control the gripper actuator. We built our own expansion board designed to power the Nucleo over the robot\u27s network and connect it to its CAN communication network. We integrated the assembled gripper as a stand-alone unit on the GLUON robot. We configured the Nucleo\u27s CPU to allow CAN communication. We based the CAN communication mode on the CANopen protocol and created dedicated messages to control the gripper. We designed and implemented a PD controller for positional control of the gripper fingers. We wrote a Python library that allows us to control the gripper from an external computer via a UDP connection. We set up a ROS environment to move the robot and taught its movements using the Pilz Industrial Motion Planner trajectory generator and demonstration files from the manufacturer. We created a Python library for robot motion using this trajectory generator. We implemented a system for teaching points by switching between the off-actuator state and the positional control of the robot. We set up the application environment, used the point teaching system to teach the robot how to execute the application, and executed it