Mobilni robotski sistem za pobiranje stebelne zelenjave

VODOPIVEC, MATIC (author), Mihelj, Matjaž (mentor), Šlajpah, Sebastjan (co-mentor)

Pričakujemo, da bosta v bližnji prihodnosti avtomatizacija in robotizacija igrali pomembno vlogo v kmetijstvu. Razloga za to sta predvsem v pomanjkanju delovne sile in v naraščanju svetovne populacije, in s tem potreb po hrani. Delo v kmetijstvu velja za fizično naporno in posledično vedno manj priljubljeno v razvitih državah. Del tega problema smo naslovili v tem magistrskem delu. Izdelali smo mobilni robotski sistem za pobiranje stebelne zelenjave. Osredotočili smo se zlasti na pobiranje špargljev. Po preučitvi obstoječih sistemov za pobiranje špargljev, smo se lotili mehanske zasnove mobilnega robotskega sistema. Robotska platforma mora biti robustna za vožnjo po neravnem terenu. Za doseganje večje produktivnosti in s tem finančne upravičenosti, mora biti pobiranje špargljev hitro in zanesljivo. Za premikanje mobilne platforme smo izbrali gosenični pogon, ki je namenjen za vožnjo po terenu. Pobiranje špargljev izvajamo s paralelnim robotom, ki je del robotskega sistema. Paralelnega robota smo izbrali zaradi majhne mase, visoke hitrosti delovanja in visoke ponovljivosti. Paralelni robot ima štiri prostostne stopnje. Tri osi poganjajo sinhroni motorji, ki so pritrjeni na bazo robota. Na platformi robota sta nameščena dva enosmerna motorja. En motor služi za rotacijo okrog vertikalne osi, drugi pa za odpiranje in zapiranje prijemala. V naslednjem koraku smo izvedli nizkonivojsko in visokonivojsko vodenje robotskega sistema. Pri nizkonivojskem vodenju smo se ukvarjali z vodenjem na nivoju krmilnika, ki je izvajal premike osi. Ukaze za premike je dobil preko komunikacije ADS iz visokonivojskega vodenja. Visokonivojsko vodenje smo implementirali v sistemu ROS. Umerili smo regulator paralelnega robota, izvedli ročno upravljanje vseh enot in načrtovali trajektorijo za pobiranje špargljev. Za zaznavo špargljev smo uporabili laserski skener, ki smo ga namestili pod kotom 45 ° glede na tla. Za izdelavo globinske slike smo uporabili pozicijsko vzorčenje. Iz zajete globinske slike je algoritem zaznal lokacije rasti špargljev. Šparglje smo sortirali po višini. Šparglji višji od 15 cm so primerni za pobiranje, nižji pa predstavljajo ovire. Delovanje mobilnega robotskega sistema smo testirali v nadzorovanem okolju. Pobiranje smo preizkusili na testnem polju špargljev. Zanesljivost pobiranja je odvisna od hitrosti platforme in gostote posajenosti špargljev.

• Language: Slovenian
• Keywords: mobilni robotski sistem, pobiranje špargljev, gosenični pogon, paralelni robot, ROS
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2022
• PID: 20.500.12556/RUL-143514
• COBISS.SI-ID: 137622019
• Publication date in RUL: 23.12.2022

Secondary language

• Language: English
• Title: Mobile robotic system for harvesting stem vegetables

Abstract

We expect automation and robotization to play an essential role in agriculture in the near future. The reasons for this are primarily the lack of a labour force, the growth of the world's population, and thus the need for food. Work in agriculture is considered physically demanding and less popular in developed countries. We addressed part of this problem in this thesis. We developed a mobile robotic system for harvesting stem vegetables. We focused primarily on picking asparagus. After studying the existing systems for picking asparagus, we specified the mechanical design of a mobile robotic system. The robot platform must be robust to travel on rough terrain. To achieve greater productivity and, thus, financial justification, harvesting asparagus must be fast and reliable. We chose a caterpillar drive designed for off-road driving to move the mobile platform. Asparagus is harvested with a parallel robot which is part of the robotic system. We chose the parallel robot because of its low mass, high operation speed and high repeatability. The parallel robot has four degrees of freedom. The three axes are driven by synchronous motors attached to the robot base. Two DC motors are mounted on the robot platform. One motor serves for rotation around the vertical axis, and the other for opening and closing the gripper. In the next step, we performed low-level and high-level control of the robot system. The low-level controller performed axis movements. It received commands for movements through ADS communication from the high-level control. We implemented high-level control in the ROS system. We tuned the controller of the parallel robot, performed manual control of all units and planned the trajectory for picking asparagus. To detect the asparagus, we used a laser scanner, which we installed at an angle of 45 ° with respect to the ground. Position-based sampling was used to create the depth image. From the captured depth image, the algorithm detected the locations of asparagus growth. We sorted the asparagus by height. Asparagus taller than 15 cm is suitable for picking, while shorter ones are obstacles. The operation of the mobile robotic system was tested in a controlled environment. We tested the harvest on a test field of asparagus. Harvesting reliability depends on the speed of the platform and the density of the asparagus growth.

• Keywords: mobile robotic system, asparagus picking, caterpillar drive, parallel robot, ROS