Development of a telerobotic test bench system for small-field-of-operation bilateral applications with 3D visual and haptic (kinaesthetic) feedback

Smit, Andre (2014-04)

Thesis (MScEng) Stellenbosch University, 2014

Thesis

ENGLISH ABSTRACT: Teleoperation as a field has seen much change since its inception in the early 1940s with Dr. Raymond Goertz producing the first teleoperation system for manipulating radioactive materials. With advances in core and supporting technologies, the systems have grown in complexity and capability, allowing users to perform tasks anywhere in the world irrespective of physical distance. The feasibility of such systems has increased as the drive for use of telepresence robots, exploration robots as in space exploration, search and rescue robots and military systems such as UAVs and UGVs gain popularity. This prompted the development of a proof of concept modular, user centred telerobotic system. The current project is the second iteration in the development process. Teleoperation and more specifically telerobotic systems pose a challenge for many system developers. This may be a result of complexity or the wide assortment of knowledge areas that developers must master in order to deliver the final system. Developers have to balance system usability, user requirements, technical design and performance requirements. Several developmental process models are considered in context of Engineering Management (EM). A larger Systems Engineering developmental process is used, with focus on the primary and supportive EM components. The author used a hybrid developmental model that is user focussed in its approach, the User-Centred Systems Design (UCSD) methodology was adopted as the primary model for application within the two distinct developmental categories. The first category hardware and system integration utilised the UCSD model as is. The second - Software development - relied on the use of agile models, rapid application development (RAD) and extreme programming (XP) were discussed with XP being chosen as it could easily incorporate UCSD principles in its development process. Hardware systems development consisted of mechanical design of end-effectors, configuration management and design, as well as haptic and visual feedback systems design for the overall physical system. Also included is the physical interface design of the input (master) cell. Further software development was broken into, three sections, the first and most important was the graphical user interface, haptic control system with kinematic model and video feedback control. The force following and matching characteristics of the system were tested and were found to show an improvement over the previous implementation. The force magnitude error at steady state was reduced by 10%. While there was a dramatic improvement in system response, the rise time was reduced by a factor 10. The system did however show a decrease in angular accuracy, which was attributed to control system limitations. Further human-factor analysis experiments were conducted to test the system in two typical use-case scenarios. The first was a planar experiment and the second a 3D placement task. The factors of interest identified were field-of-view, feedback vision mode, and input modality. Heuristic performance indicators such as time-to-completion and number of collisions for a given task were measured. System performance was only showed significant improvement when used with haptic control. This shows that the research into haptic control systems will prove to be valuable in producing usable systems. The vision factor analysis failed to yield significant results, although they were useful in the qualitative systems analysis. The feedback from post-experimentation questionnaires showed that users prefer the Point of View as a field of view and 2D viewing over 3D viewing, while the haptic input modality was preferred. The results from the technical verification process can be used in conjunction with insights gained from user preference and human-factor analysis to provide guidance for future telerobotic systems development at Stellenbosch University.

AFRIKAANSE OPSOMMING: Telewerksverigting as ’n gebied het al vele veranderinge ondergaan vandat die eerste stelsels deur Dr. Raymond Goertz geimplementeer was in die vroeë 1940s vir die hantering van radioaktiewe materiale. Met vordering in kern en ondersteunende tegnologieë, het die telewerksverigtingstelsels toegeneem in kompleksiteit asook gevorder in vermoeënsvaardigheid, wat gebruikers in staat stel om take te verrig vanuit enige plek op aarde, ongeag die fisiese afstand wat die gebruiker en die werksarea skei. Die lewensvatbaarheid van hierdie stelsels het ook toegeneem weens die belangstelling in teleteenwoordigheidrobotte, ruimtevaardige-robotte, reddings-robotte en militêre-robotte soos onbemandelug- voertuie (OLV) en onbemande-grond-voertuie(OGV). As gevolg van die belangstelling in telerobotiese stelsels is die ontwikkeling van ’n modulêre, gebruikers-gesentreerde telewerksverigting stelsel onderneem. Die huidige projek is ’n tweede iterasie hiervan. Telewerksverigting, en meer spesifiek, telerobotika stelsels ontwikelling, vereis dat stelselontwikkelaars ’n verskeidenheid kennisareas bemeester. Die ontwikkelaar moet ’n belans vind tussen gebruiker vereistes, bruikbaarheid asook tegniese ontwerp en prestasie vereistes. Menigde ontwikkelingsproses modelle is oorweeg en behandel in die konteks van Ingenieursbestuur (IB). ’n Stelselsontwikkeling proses is gevolg met ’n fokus op primêre en ondersteunende IB komponente. ’n Gemengde ontwikkeling is toegepass tot die projek wat die gebruiker as ’n hoof komponent van die stelsel in ag neem. Die oorhoofse ontwikkelingsmodel is die User-centred Systems Design (UCSD) proses, wat vir beide hardeware en sagteware ontwikkeling gebruik is. Vir die hardeware ontwikkeling is die UCSD toegepas soos dit uiteengesit is in die literatuur. Die sagteware ontwikkeling is voltooi met behulp van ratse metodes, “Rapid Application Development” RAD en “Extreme Programming” (XP) was oorweeg en XP was gekies as ontwikkelingsmodel. XP was die natuurlike keuse weens die gemak waarmee UCSD metodes en prinsiepe kon geinkorporeer word in die ontwikkelings proses. Hardeware onwikkeling het bestaan uit meganiese ontwerp, manipulasiegereedskap ontwerp, konfigurasie bestuur en ontwikkeling asook haptiese en visueleterugvoer stelselsontwerp van die fisiese stelsel insluitend die fisiese koppelvlakontwerp van die meester sel. Verder is sagtewareontwerp opgedeel in ’n haptiesebeheerstel met ’n kinematiese model ontwikkeling, videoterugvoerbeheer en gebruikersintervlak ontwerp. Die vermoëe van die stelsel om krag insette na te boots was verbeter met ’n gestadigde verbetering van 10%. Die reaksietyd van die stelsel is verbeter met ’n faktor van 10. Die stelsel het ’n verswakking getoon in die algehele hoekakkuraatheid, die oorsprong van die verswakking kan aan die beheerstelsel teogeken word. Verdere menslikefaktoranalise eksperimente is voltooi om die stelsel in twee tipiese gebruikgeval scenario’s te toets. Die eerste, ’n platvlak-eksperiment en die tweede ’n 3D plasingingstaak eksperiment. Die faktore van belang is ïdentifiseer as, visie-veld, terugvoervisie modus en insette modaliteit. Heuristiese prestasie-aanwysers soos tyd-tot-voltooiing en die aantal botsings vir ’n gegewe taak is gemeet. Stelselprestasie het slegs aansienlike verbetering getoon wanneer die stelsel met die haptiesebeheer modus bedryf word. Die visiefaktor ontleding het geen noemenswaardige resultate opgelewer nie. Terugvoervorms was na elke eksperiment voltooi. Vraelyste het getoon dat gebruikers die oogpunt van ’n lae hoek en 2D video oor 3D video verkies, terwyl die haptic beheer modaliteit verkies word.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/86516
This item appears in the following collections: