Mobilni robotski sistem za avtomatsko vzorčenje prsti

magistrsko delo

Matic Uršič (Author), Matjaž Mihelj (Mentor), Peter Kmecl (Co-mentor)

Abstract

Robotizacija v kmetijstvu predstavlja ključen korak pri napredku kmetijske dejavnosti. V magistrskem delu sta predstavljena razvoj in zgradba avtonomnega robota za vzorčenje prsti. Robot je sestavljen iz mobilne platforme, ki se premika po prostoru, ter robotskega manipulatorja, ki v želeni točki vzorči prst. Delo zajema celotno konstrukcijo mobilne platforme in manipulatorja za vzorčenje prsti. Razvili smo nizkonivojsko in visokonivojsko vodenje platforme z goseničnim pogonom, v katero smo vključili senzor globalne pozicije ter sistem za vzorčenje. Raziskovanje smo začeli s pregledom literature in teoretičnimi osnovami na področju robotike v kmetijstvu, nadaljevali pa s pregledom že obstoječih rešitev. Mobilno robotsko platformo smo konstruirali v programu za 3D modeliranje in jo kasneje v Laboratoriju za robotiko tudi sestavili. Platformi smo dodali baterijsko napajanje, ki je omogočilo avtonomijo robotskega sistema brez napajanja iz električnega omrežja. Za delovanje sistema smo uporabili tri servo-motorje Beckhoff ter dva enosmerna motorja Maxon. Za nizkonivojsko vodenje smo uporabili programsko okolje TwinCAT, z nadgradnjo pa smo vzpostavili tudi UDP komunikacijo s programskim okoljem ROS. V ROS smo z vozlišči rešili probleme navigacije, lokalizacije ter vizualizacije premikanja sistema. Vsi procesi so bili vodeni preko operacijskega sistema ROS, ki je bil nameščen na računalnik Raspberry PI 4B. S paketom Ros Navigation smo vključili globalno ter lokalno planiranje poti, ki robota vodi v željene točke. Globalni navigacijski satelitski sistem z realnočasnim kinematičnim pozicioniranjem je omogočil, da smo odometrijo, izračunano iz dajalnikov motorjev, popravili za natančnejše določanje pozicije. Virtualno okolje Rviz nam prikaže vizualizacijo robota ter okolice. Sistem za vzorčenje prsti v podanih točkah s svedrom izkoplje vzorec prsti ter ga shrani v posodo. Razvite postopke smo testirali, vrednotili ter primerjali z uporabniškimi zahtevami delovanja.

Keywords

mobilne platforme;mobilni manipulatorji;roboti;vzorčenje prsti;navigacija;lokalizacija;planiranje poti;ROS;magisteriji;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.09 - Master's Thesis
Organization:	UL FE - Faculty of Electrical Engineering
Publisher:	[M. Uršič]
UDC:	007.52:631(043.3)
COBISS:	80289539
Views:	189
Downloads:	77
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Mobile robot system for automatic soil sampling
Secondary abstract:	Robotization is one of the key steps in the progress of agriculture. The thesis addresses the development and construction of an autonomous soil sampling robot consisting of a mobile platform and a manipulator that does the soil sampling at the desired point. The thesis covers the construction of the mobile platform and the soil sampling manipulator. We developed low-level and high-level motion control of the platform, which included global positioning and a sampling system. We started the research with a literature review and theoretical foundations in agriculture robotics and continued with the analysis of existing solutions. The mobile robotic platform was designed using 3D modelling software and later assembled in the Laboratory of robotics. The battery power supply enabled the autonomy of the robotic system. Three Beckhoff servomotors and two Maxon DC motors were used to operate the system. We used the TwinCAT environment for low-level control and established UDP communication with the ROS operating system for high-level control. We solved the problems of navigation, localization and visualization of system movement with ROS nodes. All processes were programmed in the ROS operating system, installed on the Raspberry PI 4B computer. We implemented global and local path planning using the Ros Navigation package that navigates the robot to the desired points. A global navigation satellite system with real-time kinematics and combined with odometry based on the motor encoders enabled more accurate positioning. The Rviz virtual environment visualizes the robot and its surroundings. The soil sampling system with a drill takes a soil sample at a given point and stores it in a container. The developed procedures were tested, evaluated and compared with performance requirements.
Secondary keywords:	mobile platform;mobile manipulator;robot;soil sampling;navigation;localization;path planning;ROS.;
Type (COBISS):	Master's thesis/paper
Study programme:	1000316
Embargo end date (OpenAIRE):	1970-01-01
Thesis comment:	Univ. v Ljubljani, Fak. za elektrotehniko
Pages:	XXII, 60 str.
ID:	14206932