Izdelava mobilnega robota z ultrazvočnimi senzorji razdalje za lokalizacijo

Tomšič, Janez (author), Skočaj, Danijel (mentor), Rijavec, Vid (comentor)

Natančna lokalizacija je eden izmed temeljnih izzivov pri razvoju avtonomnih mobilnih robotov. Sodobni roboti za lokalizacijo običajno uporabljajo laserske merilnike razdalje (LiDAR). V diplomski nalogi raziščemo možnost uporabe cenovno ugodnejših ultrazvočnih senzorjev. Razvili smo mobilnega robota, opremljenega z dvanajstimi ultrazvočnimi senzorji HC-SR04. Za krmiljenje robota skrbi Arduino, ki komunicira s sistemom ROS na osebnem računalniku. Ovrednotili smo natančnost odometrije in adaptivne Monte Carlo lokalizacije (AMCL) v različnih znanih okoljih. Za oceno natančnosti smo implementirali sistem na podlagi sledenja značkam ArUco. Kljub omejitvam ultrazvočnih senzorjev robot dosega zadovoljivo natančnost lokalizacije z mediano napake 4,5 cm, kar je bistvena izboljšava v primerjavi z odometrijo, ki ima mediano napake 41,2 cm.

• Language: Slovenian
• Keywords: mobilna robotika, ROS, Arduino, ultrazvočni senzorji, odometrija, lokalizacija, AMCL, ArUco, 3D tiskanje
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2024
• PID: 20.500.12556/RUL-161557
• COBISS.SI-ID: 213277187
• Publication date in RUL: 12.09.2024

Secondary language

• Language: English
• Title: Development of a Mobile Robot with Ultrasonic Distance Sensors for Localization

Abstract

Precise localization is one of the fundamental challenges in the development of autonomous mobile robots. Modern robots typically use laser distance sensors (LiDAR) for localization. In this thesis, we explore the possibility of using more affordable ultrasonic sensors. We developed a mobile robot equipped with twelve HC-SR04 ultrasonic sensors. The robot is controlled by an Arduino, which communicates with the ROS system on a personal computer. We evaluated the accuracy of odometry and adaptive Monte Carlo localization (AMCL) in various known environments. To assess accuracy, we implemented a system based on ArUco marker tracking. Despite the limitations of ultrasonic sensors, the robot achieves satisfactory localization accuracy with a median error of 4.5 cm, which is a significant improvement compared to odometry, which has a median error of 41.2 cm.

• Keywords: mobile robotics, ROS, Arduino, ultrasonic sensors, odometry, localization, AMCL, ArUco, 3D printing