Thesis (MSc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: This study has started as part of a research project at Stellenbosch University (SU) that aims at building a team of soccer-playing robots for the RoboCup Small Size League (SSL). In the RoboCup SSL the Decision- Making Module (DMM) plays an important role for it makes all decisions for the robots in the team. This research focuses on the development of some parts of the DMM for the team at SU. A literature study showed that the DMM is typically developed in a hierarchical structure where basic soccer skills form the fundamental building blocks and high-level team behaviours are implemented using these basic soccer skills. The literature study also revealed that strategies in the DMM are usually developed using a hand-coded approach in the RoboCup SSL domain, i.e., a specific and fixed strategy is coded, while in other leagues a Machine Learning (ML) approach, Reinforcement Learning (RL) in particular, is widely used. This led to the following research objective of this thesis, namely to develop basic soccer skills using RL for the RoboCup Small Size League. A second objective of this research is to develop a simulation environment to facilitate the development of the DMM. A high-level simulator was developed and validated as a result. The temporal-difference value iteration algorithm with state-value functions was used for RL, along with a Multi-Layer Perceptron (MLP) as a function approximator. Two types of important soccer skills, namely shooting skills and passing skills were developed using the RL and MLP combination. Nine experiments were conducted to develop and evaluate these skills in various playing situations. The results showed that the learning was very effective, as the learning agent executed the shooting and passing tasks satisfactorily, and further refinement is thus possible. In conclusion, RL combined with MLP was successfully applied in this research to develop two important basic soccer skills for robots in the RoboCup SSL. These form a solid foundation for the development of a complete DMM along with the simulation environment established in this research. / AFRIKAANSE OPSOMMING: Hierdie studie het ontstaan as deel van 'n navorsingsprojek by Stellenbosch Universiteit wat daarop gemik was om 'n span sokkerrobotte vir die RoboCup Small Size League (SSL) te ontwikkel. Die besluitnemingsmodule (BM) speel 'n belangrike rol in die RoboCup SSL, aangesien dit besluite vir die robotte in die span maak. Hierdie navorsing fokus op ontwikkeling van enkele komponente van die BM vir die span by SU. 'n Literatuurstudie het getoon dat die BM tipies ontwikkel word volgens 'n hiërargiese struktuur waarin basiese sokkervaardighede die fundamentele boublokke vorm en hoëvlak spangedrag word dan gerealiseer deur hierdie basiese vaardighede te gebruik. Die literatuur het ook getoon dat strategieë in die BM van die RoboCup SSL domein gewoonlik ontwikkel word deur 'n hand-gekodeerde benadering, dit wil s^e, 'n baie spesifieke en vaste strategie word gekodeer, terwyl masjienleer (ML) en versterkingsleer (VL) wyd in ander ligas gebruik word. Dit het gelei tot die navorsingsdoelwit in hierdie tesis, naamlik om basiese sokkervaardighede vir robotte in die RoboCup SSL te ontwikkel. 'n Tweede doelwit was om 'n simulasie-omgewing te ontwikkel wat weer die ontwikkeling van die BM sou fasiliteer. Hierdie simulator is suksesvol ontwikkel en gevalideer. Die tydwaarde-verskil iterariewe algoritme met toestandwaarde-funksies is gebruik vir VL saam met 'n multi-laag perseptron (MLP) vir funksiebenaderings. Twee belangrike sokkervaardighede, naamlik doelskop- en aangeevaardighede is met hierdie kombinasie van VL en MLP ontwikkel. Nege eksperimente is uitgevoer om hierdie vaardighede in verskillende speelsituasies te ontwikkel en te evalueer. Volgens die resultate was die leerproses baie effektief, aangesien die leer-agent die doelskiet- en aangeetake bevredigend uitgevoer het, en verdere verfyning is dus moontlik. Die gevolgtrekking is dat VL gekombineer met MLP suksesvol toegepas is in hierdie navorsingswerk om twee belangrike, basiese sokkervaardighede vir robotte in die RoboCup SSL te ontwikkel. Dit vorm 'n sterk fondament vir die ontwikkeling van 'n volledige BM tesame met die simulasie-omgewing wat in hierdie werk daargestel is.

Thesis (MScEng)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: RoboCup promotes research in robotics and multi-robot systems (MRS). The RoboCup Small Size League (SSL), in particular, offers an entry level opportunity to take part in this field of study. This thesis presents a starting phase for research in robotics and MRS at Stellenbosch University. It includes the full documentation of the mechanical, electronic and software design of an omni-directional soccer robot for RoboCup SSL. The robot is also meant to operate as a hardware and software development platform for research in MRS. The platform was therefore designed with high-level programming language compatibility, a wide range of connectivity, and modularity in mind. The robot uses a single board computer (SBC) running a Linux operating system to accomplish these objectives. Moreover, a driver class library was written in C++ as a software application interface (API) for future development on the robot platform. The robot was also developed with a particular focus on a distributed control architecture. "Player" was implemented as the middleware, which can be used for communication between multiple robots in a distributed environment. Additionally, three tests were performed to demonstrate the functionality of the prototype: a PI speed control test, a direction accuracy test and a static communication test using the middleware. Recommendations for possible future work are also given. / AFRIKAANSE OPSOMMING: RoboCup bevorder navorsing in robotika en multi-robot-stelsels (MRS). Die RoboCup Klein Liga (KL) bied in die besonder die geleentheid om op intreevlak navorsing te doen in hierdie veld. Hierdie tesis verteenwoordig die eerste fase van navorsing in robotika en MRS by Stellenbosch Universiteit. Dit sluit die volledige dokumentasie van die meganiese, elektroniese en sagteware-ontwerp van ’n omnidireksionele sokker-robot vir die KL in. Die robot is ook veronderstel om te dien as ’n hardeware- en sagteware-ontwikkelingsplatform vir navorsing in MRS. Die platform is dus ontwerp met ’n verskeidenheid van uitbreingsmoontlikhede en modulariteit in gedagte asook die moontlikheid om gebruik te maak van ’n hoë-vlak programmeertaal. Om hierdie doelwitte te bereik, maak die robot gebruik van ’n enkel-bord-rekenaar met ’n Linux bedryfstelsel. Verder was ’n sagteware drywer in C++ geskryf om te dien as ’n sagteware-koppelvlak vir toekomstige ontwikkeling op die robot platform. Die robot is ook ontwikkel met die besondere fokus op ’n gedesentraliseerde beheerstels. Player was geïmplementeer as die middelware, wat gebruik kan word vir kommunikasie tussen verskeie robotte in ’n gedesentralliseerde beheerstelsel. Daar is drie toetse uitgevoer om die funksionaliteit van die prototipe te demonstreer, ’n PI spoed beheer toets, ’n rigting akkuraatheidstoets en ’n statiese kommunikasie toets deur van die middelware gebruik te maak. Aanbevelings vir moontlike toekomstige werk word ook verskaf.

Thesis (MEng)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: The RoboCup offers a research platform to advance robotics and multi-robot cooperation in dynamic environments. This project builds on work previously done to develop a research platform for multi-robot cooperation at Stellenbosch University. This thesis describes the development of an active ball handling system for a robot in the RoboCup Small Size League (SSL). This was achieved by building on the work done in the previous projects. The hardware for the kicker and dribbler mechanisms on the robot were implemented and tested to characterise their capabilities. The kicker was characterised to control the speed at which a ball is kicked and the dribbler for optimal control over a ball. More accurate movement was required and the Proportional Integral and Derivative (PID) controllers for translational and rotational movement on the robot were improved. The test results show an improvement in straight line trajectory tracking when compared to those of the previous controllers. Dribble control sensors were implemented on the robot for successful dribbling by the robot. This resulted in a significant improvement to the dribbling ability of the robot when these sensors are used. This dribbling ability was compared to the dribbling ability of the robot when no feedback was received from the sensors. Lastly a proposed curved trajectory tracking algorithm was tested by combining translational and rotational movement of the robot. This algorithm showed the capabilities of the robot to follow a curved trajectory with the improved translational and rotational controllers. / AFRIKAANSE OPSOMMING: Die RoboCup bied ’n navorsingsplatvorm om robotika en multi-robot samewerking in ’n dinamiese omgewing te bevorder. Hierdie projek bou voort op werk wat reeds gedoen is om ’n navorsingsplatvorm vir multi-robot samewerking aan die Universiteit van Stellenbosch te ontwikkel. Hierdie tesis beskryf die ontwikkeling van ’n aktiewe balhanteringsstelsel vir ’n robot in die RoboCup Klein Liga (KL). Dit is bereik deur voort te bou op die werk wat in vorige projekte gedoen is. Die hardeware vir die skopper- en dribbelmeganismes is geïmplementeer en getoets om hulle vermoëns te karakteriseer. Die skopper is gekenmerk deur die spoed waarteen ’n bal geskop word en die dribbler vir optimale beheer oor ’n bal. Meer akkurate beweging was nodig en die PID-beheerders vir translasie- en rotasiebeweging in die robot is verbeter. Die resultate van die toetse toon ’n verbetering in reguitlynbeweging in vergelyking met dié van die vorige beheerders. Dribbelbeheersensors is in die robot geïmplementeer vir suksesvolle dribbelbeweging deur die robot. Gevolglik is daar ’n aansienlike verbetering in die dribbelvermoë van die robot wanneer hierdie sensors gebruik word. Hierdie dribbelvermoë is vergelyk met die dribbelvermoë wanneer die robot geen terugvoer van die sensors ontvang nie. Laastens is ’n voorgestelde algoritme vir die robot om ’n geboë trajek te volg, getoets. Dit is bereik deur die translasie- en die rotasiebeweging van die robot te kombineer. Hierdie algoritme het die vermoë van die robot om ’n geboë baan te laat volg deur gebruik te maak van die verbeterde translasie- en rotasiebeheerders.