AAU Student Projects - visit Aalborg University's student projects portal
A master's thesis from Aalborg University
Book cover


Real World Development and Test of a Mobile Hybrid Robot Platform.

Authors

;

Term

4. semester

Education

Publication year

2022

Submitted on

Pages

86

Abstract

Dette projekt undersøger, hvordan man designer en hybridrobot, der kan skifte mellem at køre som en bil og balancere som en Segway. Skiftet sker via en sving-op-manøvre, der løfter robotten og holder den oprejst i en omvendt pendul-tilstand—som at balancere en kost lodret på hånden. For at afprøve ideer sikkert og billigt blev der udviklet en simulering, så design og funktioner kan testes før afprøvning i den virkelige verden. Simuleringen blev også brugt til at estimere de nødvendige motoregenskaber under hensyn til pris og tilgængelighed. Test viser, at et sving-op kan gennemføres efter opnåelse af en vis hastighed, og at oprejst balance kan stabiliseres ved hjælp af forskellige styringsmetoder, herunder PID-regulering, simpel tilstandsfeedback og LQR. Arbejdet er et lovende første skridt, men konstruktion og valg af komponenter kræver yderligere forfinelse og optimering, før robotten kan fungere optimalt i virkelige anvendelser.

This project explores how to design a hybrid robot that can switch between driving like a car and balancing like a Segway. The transition uses a swing-up maneuver that lifts the robot and then holds it upright in an inverted-pendulum state—like balancing a broomstick on your hand. To test ideas safely and affordably, a simulation was developed to evaluate designs and features before real-world trials. The simulation also helped estimate the required motor characteristics within budget and availability limits. Tests indicate that a swing-up is achievable once a certain speed is reached, and that upright balancing can be stabilized using control methods such as PID control, simple state feedback, and LQR. This is a promising first step, but the design and component selection need further refinement and optimization before the robot can operate reliably in practical, real-world settings.

[This summary has been rewritten with the help of AI based on the project's original abstract]