1,721,020 research outputs found

    Modeling and Control of an Autonomous Racing Car

    No full text
    Denne afhandling beskriver design og analyse af en ikke-lineær model og en kontrolstrategi for Aalborg University Race teamets G8 racerbil og er motiveret af behovet for en Autonomous Racing Car (ARC) til at konkurrere i Formula Student-arrangementet i 2021. Præstationen af den designede model evalueres i et simuleringsmiljø implementeret i Matlab Simulink. Der er gjort en indsats for at holde det genanvendeligt til udvidelser i det fremtidige arbejde. Bevægelsesligningerne er formuleret, inklusive dækdynamik sammen med aerodynamisk forstyrrelse og tager hensyn til langsgående og lateral dynamik. En linearisering af den ikke-lineære model tillader en implementering af en linear quadratic controller (LQR) design, til at regulere hastigheden. Modellen viser god ydelse i regioner over 36 kilometer i timen, men regionen under lider af en forenklet motormodel.This thesis describes the design and analysis of an nonlinear model and a control strategy for Aalborg University Race team’s G8 race car and has been motivated by the need for an Autonomous Racing Car (ARC) to compete in the Formula Student event in 2021. The performance of the designed model is evaluated in a simulation environment, implemented in Matlab Simulink. Effort has been put in to keep it reusable for extensions in future work. The equations of motion have been formulated, including tire dynamic together with aerodynamic disturbance and take into account longitudinal and lateral slip. A linearization of the nonlinear model has be performed to allow an implementation of a linear quadratic controller (LQR) design, to regulate the speed. The model shows good performance in regions above 36 kilometer per hour, but region below suffer from a simplified engine model

    Formation Control for a CubeSat Constellation

    No full text
    This report introduces formation flying control strategies for CubeSat's Constellations in low Earth orbit.The use of CubeSats for short missions has increased exponentially due to the versatility and relative low costs of these type of spacecrafts.The main disadvantage of CubeSats is the reduced space to store equipment, thus some missions use from two to twenty four CubSats to achieve better results, this configurations are known as CubeSat's constellations.Some common applications are the measurement of magnetic fields and inspection of larger spacecraft's surroundings for maintenance and repair purposes, therefore the problem statement can be set as follow:"Design a distributed control strategy for a constellation of CubeSats, where the satellites fulfill separation and velocity constraints between them and a target."This project will discuss the station keeping dynamics and control of CubSats as well as control techniques for formation flying of Cubesat's constellations orbiting Earth

    Satellite Attitude Control: Using Magnetorquers with Magnetic Dipole Moment Calcellation

    No full text
    This thesis describes the design and anal- ysis of a nonlinear attitude control system for a CubeSat sized satellite based on the structure of the AAUSAT satellites.An analysis of the main disturbances is made with the conclusion that only the magnetic dipole moment needs to be taken into account when performing at- titude control, as this disturbance is 10 times larger than the rest combined.An estimation of the satellite dipole mo- ment is made as a bias estimation by an Extended Kalman Filter. The resulting estimate is correct within 8 % of the true dipole moment when estimating on data basen on a slowly tumbling satellite with state noise.A sliding mode controller is used as the nonlinear controller and uses the mag- netic dipole estimate from the Extended Kalman Filter to counteract this distur- bance. The result is a controller capa- ble of following a Nadir reference within an axis-wise error of 10 degrees. The downside to the controller is the constant power consumption needed, in order to counteract the magnetic dipole moment of the satellite

    Formation Control of Autonomous Surface Vehicles for Surveying Purposes

    No full text
    Dette kandidatspeciale omhandler fortsat arbejde på AAUSHIP platformen. AAUSHIP er en ASV som kan have flere forskellige anvendelsesmuligheder. Fokus i dette projekt er formationskontrol med baggrund i pejlingsopgaver hvortil AAUSHIP skal udvides fra et enkelt skib til en flåde af AAUSHIPs.Båden er fysisk blevet opgraderet med nyt elektronik som er implementeret. Dertil er den enkelte båd blevet testet med et nyt KF hvortil en retningsregulator også er implementeret. Denne er anvendt som en LOS reference regulator for at få båden til at konvergere til de genererede liniestykker mellem rutepunkter. Efterfølgende har fokus været sat på at identificere og analysere formationsstrategier som skal implementeres på bådene, når disse er produceret. Mest fokus er lagt på en potentialefeltsalgoritme, som er blevet simuleret inklusiv dynamikken fra modellen af AAUSHIP. Dette danner grundlaget for implementering på flåden med opfølgende verifikation.Resultater viser, at det er muligt at kontrollere AAUSHIP med den designede model, i et område givet af Aalborg Havn. Yderligere arbejde vil ligge i at forbedre modellen af AAUSHIP, men dette har ikke vist sig at være nødvendigt da fokus i dette project har omhandlet pejlingsopgaven. Slutteligt udviser simulering af formationskontrolstrategien potentiale for en mulig implementering af denne på den kommende flåde af AAUSHIPs, så disse kan foretage pejligsopgaver som en dannet formation af skibe.This master's thesis concerns the further work on the platform named AAUSHIP. This is an ASV which can have different purposes. Within the scope of this project it will be used for surveying applications, where the AAUSHIP will be expanded to be a fleet of AAUSHIPs to navigate in a formation.Firstly, the old AAUSHIP is upgraded with respect of hardware and implementation of these have been necessary. The single AAUSHIP have been tested after newly implemented Kalman Filter and a heading controller have been implemented. This is used as a LOS guidance to make the AAUSHIP converge onto a predetermined trajectory. Afterwards is the focus to investigate formation strategies to be implemented at the AAUSHIP when more ships are to come. The main investigated strategy is based on a potential field algorithm, which have been simulated with the dynamics of the AAUSHIP. This is the basis of future work to implement this strategy at the coming AAUSHIP fleet for verification of the methods.Results show that, with the model designed, it is possible to control the AAUSHIP in the area of interest. Further work to the model can improve performance, but this has not been the main focus within the scope of this project, where the surveying purpose have been in focus. Simulation of the formation control strategy with potential field shows the potential to implement this at the coming AAUSHIP fleet, such that these will be able to perform surveying as an entire group of vessels

    Navigation for Autonomous Surface Vessels

    No full text
    Denne afhandling omhandler derrivation af en ulineær model for en autonom båd. Med denne model udvikles simulation, kontrolsystemer og to navigationsalgoritmer.This thesis covers derivation of a non-linear model, simulation of this model, development of control systems and path planning systems for an Autonomous Surface Vessel. A nonlinear dynamic model was created using Newtonian mechanics in combination with hydrostatics. This model was implemented in both Simulink and Gazebo. A PID controller was implemented, and the possibility of applying LPV control was investigated. The Artificial Potential Field method and the State Lattice method were both implemented. Measurements made on the model were used to ensure that the State lattice was restricted to only finding paths that the system is capable of following.<br/

    Phasing of Satellites in a Single Orbital Plane: A Lyapunov Control Approach

    No full text
    The aim of this project was to phase grouped satellites in a single orbital plane. The project was done as a part of the research project MARIOT, which seeks to develop a satellite based maritime IoT network using VDES communication. First, a problem analysis investigating the different aspects of a satellite in orbit are presented. This included orbit parameters, different actuation methods, specifications of antennas, and some satellite constellations currently providing communication coverage around the world. Kinematics and dynamics describing a satellite with respect to another satellite in orbit were defined. The dynamics included modelling of gravitational and atmospheric drag perturbations, which was verified using the program AGI STK. A PD-, LQR-, and Lyapunov controller was implemented and tested in MATLAB/Simulink. The PD controller was stable distant to the reference, the LQR controller was stable near the reference, and the Lyapunov controller was stable all the way. The Lyapunov control was introduced to drag and remained stable. The implementation of the system was deemed successful

    State Estimation and Control of a Self-Foraging Rocket

    No full text
    This master thesis describes the design anddevelopment of position and attitude determinationand control of a rocket. The projectis motivated by the general need of cheaperlaunch opportunities for an ever growing microsatellite market. The goal is to achieve acheap and efficient solution in the form of theself-foraging rocket concept, inspired by thework of Yemets et al. at Dniepropetrovsk NationalUniversity.The overall design of the rocket is done fromscratch starting with the initial mechanicalconcept of the rocket, followed by the sensorsand actuators it will employ, and consecutivelythe mechanical dimensioning of eachthruster to be utilized. Additionally a generalascent trajectory is calculated to launch therocket into orbit, to be tested in a completesimulation environment designed in Simulink.Furthermore an extended Kalman filter is designedto estimate the states of the rocket,while both waypoint and sliding mode is designedfor the control of the rocket.The EKF was shown to be able to sufficientlyestimate the states of the system in three distinctnominal tests.A waypoint tracking controller, for followingan optimal ascent trajectory, and a slidingmode controller is implemented as positioncontrol systems for launching the rocket intoorbit. Despite for showing good performancefor controlling the rocket, none of the controllersare able to inject the rocket into orbitwith satisfactory resultsThis master thesis describes the design anddevelopment of position and attitude determinationand control of a rocket. The projectis motivated by the general need of cheaperlaunch opportunities for an ever growing microsatellite market. The goal is to achieve acheap and efficient solution in the form of theself-foraging rocket concept, inspired by thework of Yemets et al. at Dniepropetrovsk NationalUniversity.The overall design of the rocket is done fromscratch starting with the initial mechanicalconcept of the rocket, followed by the sensorsand actuators it will employ, and consecutivelythe mechanical dimensioning of eachthruster to be utilized. Additionally a generalascent trajectory is calculated to launch therocket into orbit, to be tested in a completesimulation environment designed in Simulink.Furthermore an extended Kalman filter is designedto estimate the states of the rocket,while both waypoint and sliding mode is designedfor the control of the rocket.The EKF was shown to be able to sufficientlyestimate the states of the system in three distinctnominal tests.A waypoint tracking controller, for followingan optimal ascent trajectory, and a slidingmode controller is implemented as positioncontrol systems for launching the rocket intoorbit. Despite for showing good performancefor controlling the rocket, none of the controllersare able to inject the rocket into orbitwith satisfactory result
    corecore