Modelsko prediktivno upravljanje mobilnim robotom

master's thesis

• Title (english): Model predictive control of a mobile robot
• Author: Jurica Vučković
• Mentor: Andrej Jokić (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2024-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Područje upravljanja mobilnim robotima je u zadnjem desetljeću u povećanom fokusu znanstvene zajednice i industrije. Usprkos naizgled jednostavnom kinematskom modelu diferencijalnog mobilnog robota, postojanje neholonomskih ograničenja pretvara dizajniranje stabilizirajućih upravljačkih zakona za ovakav sustav u izazovan i netrivijalan problem. U primjenama upravljačkih algoritama na stvarnim robotskim sustavima, jedan od značajnijih izazova u sintezi regulatora čine uvijek prisutna ograničenja na ulaze i stanja sustava koja se prirodno javljaju zbog mehaničkih i fizičkih karakteristika. Ova ograničenja, ako se prikladno ne uzmu u obzir, mogu značajno narušiti ponašanje upravljanog sustava, a ponekad uzrokovati i nestabilnost zatvorenog kruga.

U ovom diplomskom radu koristi se optimalna metoda upravljanja zvana modelsko prediktivno upravljanje (engl. Model Predictive Control - MPC) za upravljanje diferencijalnim mobilnim robotom ASTRO (engl. Autonomous System for Teaching Robotics). MPC u svojoj formulaciji omogućava eksplicitno uračunavanje ograničenja na varijable stanja i ulaze sustava. Za mobilne robote mogućnost definiranja ograničenja je potrebna i za postavljanje sigurnog radnog prostora robota, osobito ako je mobilni robot namijenjen za rad u okolini s čovjekom.

Budući da je model ASTRO robota nelinearan, u ovom diplomskom zadatku implementirana je inačica nelinearnog modelskog prediktivnog upravljanja (engl. Nonlinear Model Predictive Control - NMPC) u radnom okviru robotskog operativnog sustava (engl. Robot Operating System - ROS 2). Za rješavanje nelinearnog optimizacijskog problema korištena je C++ biblioteka otvorenog koda zvana CasADi.

Abstract (english)

The field of mobile robot control has been in increased focus over the last decade due to growing interest from industry and the scientific community. Despite the seemingly simple kinematic model of a differential wheeled robot, the presence of nonholonomic constraints makes designing stabilizing control laws for such systems a challenging and non-trivial problem. In applications of control algorithms in real robotic systems, it is difficult to achieve good control performance due to the requirement of subjecting the system's states and inputs to a set of constraints that naturally arise from its mechanical and physical characteristics. If this constraints aren't adequately taken into account they can significantly impact the behavior of the control system, and can sometimes cause instability of closed-loop control.

In this master's thesis, an optimal control method called model predictive control (MPC) is used for controlling the differential wheeled robot ASTRO (Autonomous System for Teaching Robotics). Inherent to MPC's formulation is the ability to handle constraints on the system's inputs and states. The ability to handle constraints is especially important for mobile robots working nearby humans, so a safety work perimeter for the robot can be set.

Since the mathematical model of ASTRO is nonlinear, in this master's thesis, a version of nonlinear model predictive control (NMPC) has been implemented in the robot operating system (ROS 2) framework. To solve the nonlinear optimization problem, an open-source C++ library called CasADi has been used.

• Keywords: ROS 2; MPC; NMPC; optimalno upravljanje; CasADi; diferencijalni mobilni robot; ASTRO
• Keywords (english): ROS 2; MPC; NMPC; optimal control; CasADi; differential wheeled robot; ASTRO
• Language: croatian
• URN:NBN: urn:nbn:hr:235:701213
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-07-10 13:11:58