No description available.

Thesis (PhD (Electrical and Electronic Engineering))--University of Stellenbosch, 2010. / Dissertation presented in fulfilment of the requirements for the degree of Doctor of Philosophy in Engineering at Stellenbosch University. / ENGLISH ABSTRACT: Airworthiness certification of unmanned aircraft systems (UAS) is normally considered to be a regulatory function. In the absence of comprehensive UAS airworthiness regulations, the development of new and unique UAS, and their introduction into non-segregated airspace, remain major challenges for the UAS industry and regulators. Thus, in response, the objective of this research was to establish a framework and guidelines, within the scope of the typical regulatory regime, that can be used by the UAS engineering domain to ensure the safe and reliable functioning of a UAS, whether regulated or not. UAS airworthiness is currently mainly based on manned aircraft regulations, and the focus is on the unmanned aircraft and the 'airworthiness' of the remote control station. The typical UAS as a system, however, consists of more than just these elements and a broader approach to the 'airworthiness' of a UAS is required. This study investigated and introduces the concept of UAS operability, where the term 'operability' addresses the safe and reliable functioning of the UAS as a system, the airworthiness of its airborne sub-systems, and the safe and reliable functioning of its non-airborne subsystems and functional payloads. To ensure that the results of this study are aligned with typical aviation regulatory systems, a regulatory basis was defined within which UAS operability guidelines could be developed. Based on the operability concept, and in the scope of the regulatory basis, a UAS operability framework was developed for the UAS engineering domain. This framework is an index and reference source from which appropriate operability elements can be selected for a particular UAS. The scope of the framework is generic, rather than UAS-type or -class specific, and includes operability elements for the UAS as a system, for its airborne and non-airborne sub-systems, and for its payloads. The framework was validated by developing lower hierarchical levels for the framework and by populating each operability element of the framework with appropriate engineering guidance criteria. The guidance criteria were derived and/or developed from industry 'best practices' found in the literature, or were newly developed where no existing practices were found. The significance of this study is found in its establishing of a generic UAS operability framework that not only focuses on the airworthiness of the unmanned aircraft, but addresses the operability of the UAS as a system, as well as the operability of its airborne sub-systems, its non-airborne sub-systems and its payloads. In practice, the UAS operability framework can be used in the UAS engineering domain as an index and reference source to select relevant operability elements for a particular UAS. The guidance criteria for the selected elements can subsequently be used to develop the appropriate processes, procedures, requirements and specifications to achieve initial operability of the UAS, and to maintain its continued operability. Although the objective of the research was achieved, the UAS operability framework must still be applied and tested in real-life UAS projects and, where necessary, revised to eliminate shortcomings and to provide for new and novel developments in UAS engineering technologies. / AFRIKAANSE OPSOMMING: Die lugwaardigheidsertifisering van onbemande vliegtuigstelsels (OVS) word normaalweg beskou as 'n reguleringsfunksie. In die afwesigheid van omvattende OVS lugwaardigheidsregulasies bly die ontwikkeling van nuwe en unieke OVS, en die inbedryfstelling daarvan in onafgesonderde lugruim, besonderse uitdagings vir beide die OVS nywerheid en reguleerders. Die doelwit van hierdie navorsing was dus om riglyne binne die bestek van die tipiese reguleringsregime te vestig wat deur die OVS ingenieursdomein benut kan word om die veilige en betroubare funksionering van 'n OVS te verseker, of dit gereguleer word aldan nie. OVS lugwaardigheid word tans hoofsaaklik gebaseer op lugwaardigheidsvereistes vir bemande vliegtuie. Die fokus is dan ook meerendeels op die onbemande vliegtuig en die 'lugwaardigheid' van die afstandbeheerstasie. Die tipiese OVS bestaan egter uit meer sub-stelsels en 'n weier beskouing van die 'lugwaardigheid' van 'n OVS is nodig. Die konsep van OVS bedryfbaarheid is in hierdie studie ondersoek en voorgestel. 'Bedryfbaarheid' beteken in hierdie konteks die veilige en betroubare funksionering van die OVS as 'n stelsel, die lugwaardigheid van die lug sub-stelsels, die veilige en betroubare funksionering van die nie-lug sub-stelsels, asook die veilige en betroubare funksionering van funksionele loonvragte. Om te verseker dat die resultate van hierdie studie versoenbaar is met tipiese lugvaart reguleringstelsels, is 'n reguleringsbasis omskryf vir die ontwikkeling van OVS bedryfbaarheidsriglyne. Gebaseer op die bedryfbaarheidskonsep, en binne die riglyne van die reguleringsbasis, is 'n OVS bedryfbaarheidsraamwerk ontwikkel vir die OVS ingenieursdomein. Die raamwerk is 'n indeks en verwysingsbron waaruit gepaste bedryfbaarheids-elemente gekies kan word vir 'n bepaalde OVS. Die bestek van die raamwerk is generies en nie beperk tot spesifieke OVS tipes of klasse nie. Die raamwerk sluit bedryfbaarheidselemente in vir die OVS as stelsel, asook vir die lug en nie-lug sub-stelsels van die OVS, en vir die loonvragte van die OVS. Die raamwerk se geldigheid was bevestig deur die struktuur van die raamwerk tot laer vlakke uit te brei en gepaste ingenieursriglyne vir elke bedryfbaarheids-element in die raamwerk te ontwikkel. Die riglyne was gebaseer op 'beste praktyke' soos beskryf in die literatuur, of was van nuuts af ontwikkel waar geen bestaande praktyke gevind kon word nie. Die bydrae van hierdie studie is gesetel in die vestiging van 'n generiese OVS bedryfbaarheidsraamwerk wat nie net gemik is op die lugwaardigheid van die onbemande vliegtuig nie, maar wat die bedryfbaarheid in geheel van die OVS as stelsel aanspreek, asook die bedryfbaarheid van die OVS se lug sub-stelsels, nie-lug sub-stelsels en loonvragte. In die praktyk kan die raamwerk in die OVS ingenieursdomein gebruik word om gepaste bedryfbaarheids-elemente vir 'n OVS te kies. Daarna kan die bedryfbaarheidsriglyne gebruik word om gepaste prosesse, prosedures, vereistes en spesifikasies te ontwikkel om die OVS se aanvanklike en voortgesette bedryfbaarheid te bewerkstellig. Alhoewel die doelwit vir die navorsing bereik is, moet die OVS bedryfbaarheidsraamwerk nog op werklike OVS projekte getoets word. Waar nodig, moet die raamwerk dan hersien word om tekortkominge, asook nuwe en unieke ontwikkelinge in OVS ingenieurstegnologie, aan te spreek.

Unmanned air vehicles have been increasing in their autonomous capabilities. This research furthers these capabilities by focusing on the automation of landing site determination for rotorcraft in urban environments. Automated landing saves energy and allows the aircraft to choose areas that are safe for people and the aircraft. Two methods are used to gather information about the terrain of potential landing sites. One method is 3D reconstruction from multiple camera images. The other method uses a range sensor to reconstruct the terrain. Both of these methods create an inertial terrain map of the environment in the form of a point cloud that can be investigated for possible landing sites. Two strategies were developed to search the terrain map for possible landing sites: grid-based RANSAC and Recursive-RANSAC (R-RANSAC). Both strategies search for flat stable areas for landing. Grid-based RANSAC separates the terrain map into discrete portions for plane fitting analysis. These fitted planes are used to determine whether portions of the terrain map are suitable for landing. Two additional variations of grid-based RANSAC were explored that resulted in improvements to the approach. This strategy can quickly find landing sites from large terrain maps. The other strategy, R-RANSAC, is a recursive approach that analyzes each point in the terrain map for plane fitting. New planes are created as needed to fit points in the terrain map. Planes that fit a large number of points are analyzed for possible landing locations. This strategy is more complex to implement, but results in a simpler model of the environment: a small set of 3D planes. The results are displayed with the possible landing locations. Both landing-site strategies were implemented onboard a hexrotor aircraft and successfully demonstrated in flight.

Wind observations that are critical for understanding meteorological processes occurring inside of the Earth's atmospheric boundary layer (ABL) are sparse due to limitations of conventional atmospheric sensors. In this dissertation, dynamic systems and estimation theory are combined with experimental methods to exploit the flight envelope of multirotor UAS for wind sensing. The parameters of three quadrotor motion models, consisting of a kinematic particle, a dynamic particle, and a dynamic rigid body models are developed to measure wind velocity in hovering flight. Wind tunnel and steady level flight tests are used to characterize kinematic and dynamic particle models. System identification stepwise regression and output error algorithms are used to determine the model structure and parameter estimates of rigid body models. The comparison of all three models demonstrates the rigid body model to have higher performance resolving slow-varying winds based on a frequency response analysis and field experiments conducted next to a 3-D sonic anemometer. The dissertation also presents an extension of the rigid body wind estimation framework to profile the horizontal components of wind velocity in vertical steady ascending flight. The extension employed system identification to characterize five rigid body models for steady-ascending flight speeds increasing from 0 to 2 m/s in intervals of 0.5~m/s. State observers for wind profiling were synthesized using all five rigid body models. Performance assessments employing wind observations from in situ and remote sensors demonstrated model-based wind profiling results to be be in close agreement with ground-truth wind observations. Finally, the rigid body wind sensing framework developed in this dissertations for multirotor UAS is employed to support science objectives for the Advanced Lagrangian Predictions for Hazards Assessment Project. Quadrotor wind measurements sampled at 10 m above sea level were used to characterize the leeway of a person in water for search and rescue scenarios. Leeway values determined from quadrotor wind measurements were found to be in close to leeway parameters previous published in the literature. This results demonstrates the utility of model-based wind sensing for multirotor UAS for providing wind velocity observations in complex environments where conventional wind observations are not readily available. / Doctor of Philosophy / Wind observations that are critical for understanding meteorological processes occurring inside of the Earth's atmospheric boundary layer (ABL) are sparse due to limitations of conventional atmospheric sensors. In this dissertation, dynamic systems and estimation theory are combined with experimental methods to exploit the flight envelope of multirotor UAS for wind sensing. The parameters of three quadrotor motion models, consisting of a kinematic particle model, a dynamic particle model, and a dynamic rigid body model, are characterized to measure wind velocity in hovering flight. Parameter characterizations are realized using data from wind tunnel, steady level flight tests and system identification experiments. Model-based wind estimations algorithms are developed using the kinematic particle model directly and by synthesizing state observers for the dynamic particle and rigid body models separately. For comparison purposes, the frequency response characteristic of the dynamic particle and rigid body models is examined to determine the range of wind fluctuations that each model can resolve. Performance comparisons demonstrate that the rigid body model to resolve higher wind fluctuations and yield more accurate wind estimates. The dissertation extends the rigid body wind estimation algorithm to estimate wind velocity profiles of the horizontal wind vector. The rigid body wind estimation algorithms is used to answer science questions about about the drift of a person in water.

This dissertation addresses improvements to a consensus based task allocation algorithms for improving the Quality of Service in multi-task and multi-agent environments. Research in the past has led to many centralized task allocation algorithms where a central computation unit is calculating the global optimum task allocation solution. The centralized algorithms are plagued by creating a single point of failure and the bandwidth needed for creating consistent and accurate situational awareness off all agents. This work will extend upon a widely researched decentralized task assignment algorithm based on the consensus principle. Although many extensions have led to improvements of the original algorithm, there is still much opportunity for improvement in providing sufficient and reliable task assignments in real-world dynamic conditions and changing environments. This research addresses practical changes made to the consensus based task allocation algorithms for improving the Quality of Service in multi-task and multi-agent environments.

The United States Air Force has specified a need for the next generation, High Altitude, Long Endurance aircraft capable of carrying advanced sensor arrays over very large distances and at extreme altitudes. These extensive set of requirements has required a radical shift away from the conventional wing & tube configurations with a new focus placed on extremely light weight and unconventional structural and aerodynamic configurations. One such example is the Boeing Joined wing SensorCraft Concept. The Joined wing concept has potential structural and sensor carrying benefits, but along with these potential benefits come several challenges. One of the primary concerns is the aeroelastic response of the aft wing, with potential adverse behaviours such as flutter and highly nonlinear structural behaviour of the aft wing under gust conditions. While nonlinear computation models have been developed to predict these responses, there exists a lack of experimental ground and flight test data for this unique joined wing configuration with which to benchmark the analytical predictions. The goal of this work is to develop a 5m, scaled version of the Boeing Joined Wing configuration and collect data, through a series of ground and flight based tests, which will allow designers to better understand the unique structural response of the configuration. A computational framework was developed that is capable of linearly scaling the aeroelastic response of the full scale aircraft and optimize a reduced scale aircraft to exhibit equivalent scaled behaviour. A series of reduced complexity models was developed to further investigate the flying characteristics of the configuration, test avionics and instrumentation systems and the develop flight control laws to adequately control the marginally stable aircraft. Lessons learned were then applied the 5m flight test article that was designed and constructed by the author. In the final stage of the project, the decision was made to relax the aeroelastically scaled constraint in order to allow additional softening of the structure to further investigate the nonlinear behaviour of the aircraft. Due to the added risk and complexity of flying this highly flexible aircraft the decision was made to produce the final aeroelastically scaled article at the 1.85m scale. This model was designed, developed and ground tested in the lead up to a follow on project which will see additional flight testing performed in conjunction with Boeing Inc. / Graduate

Over the last years there has been a rapid development in the UAS-technology (Unmanned Aircraft Systems) and today there are several UAS systems on the market. The fast development has led to differences in both price and capability of taking high-quality images between the systems. The purpose of this study was firstly to investigate how two UAS systems differ in the uncertainty of measurement while making digital terrain models, secondly, to investigate how different UAS systems cope with the laws and requirements that exist for producing digital terrain models for detail projection, SIS-TS 21144:2016 Table 6 level 1-3. A comparative study on two software’s creation of point clouds from picture data was also conducted. In this study, three digital models were made from one specific area. They were created with two different UAS-systems and laser scanning from an airplane. The models were compared and analysed using the RUFRIS method. The UASsystems used were a fixed wings Smartplanes S1C and a rotary wings Dji Phantom 4 PRO. The Smartplanes flew 174 m above the ground and the Dji Phantom 4 flew 80 m above the ground. The results from the study show that laser scanning from the airplane created the model with the lowest measurement uncertainty and met all the requirements for each separate type (asphalt, natural soil, grass and gravel) for detail projection according to SIS.TS 201144:2016 table 6 level 1-3. Additionally, the results show that the terrain model produced by the Dji Phantom 4 only met the requirements for asphalt where the mean deviation was 0,001 m. The results produced with “Smartplanes” met the requirements for asphalt and gravel where the mean deviations were -0,007 m and 0,017 m. The softwares PhotoScan and UASMaster were compared while creating point clouds from pictures taken by the Smartplanes. The results show that PhotoScan had the lowest uncertainty for asphalt, grass and gravel surfaces while UASMaster produced lower uncertainty for natural soil. The results indicate that airborne laser scanning should be the preferred method for collection of topographic data since it created lower measurement uncertainties than the other methods in this study. It is also possible to create digital terrain models with UAS for detail projection for asphalt and gravel surface in accordance with 21144:2016. Finally, it was concluded that the used software programs are showing differences in creating point clouds. / De senaste åren har tekniken för Unmanned Aircraft System (UAS) utvecklats snabbt och idag finns flera system på marknaden. Ett resultat av den snabba utvecklingen är att de olika systemen skiljer sig åt, dels i pris men även i kapacitet. Syftet med studien var att undersöka hur olika UAS-system skiljer sig åt i mätosäkerhet vid framställning av digitala terrängmodeller, men även hur olika UAS-system står sig mot det regelverk som finns för framställning av digitala terrängmodeller vid detaljprojektering enligt SIS-TS 21144:2016 Tabell 6 klass 1-3. Ytterligare ett syfte med studien var att undersöka hur olika programvaror skiljer sig åt vid framställning av punktmoln från bilddata. I studien kontrollerades och jämfördes tre digitala terrängmodeller genererade över samma område med två olika UAS-system samt laserskanning från ett flygplan. Terrängmodellerna jämfördes mot kontrollprofiler framställda med RUFRIS-metoden. De olika UAS-systemen var en dyrare variant, Smartplanes S1C (fastavingar), och en billigare variant, Dji Phantom 4 PRO (roterande vingar). De tillämpade flyghöjderna för flygningarna var 174 m för Smartplanes och 80 m för Dji Phantom. Resultatet från studien visar att laserskanning från flygplanet uppnådde lägst mätosäkerhet och klarade samtliga krav för varje separat marktyp för detaljprojektering enligt SIS-TS 201144:2016 Tabell 6 klass 1-3. Marktyper som undersöktes var: asfalt, naturmark, gräs och grus. Vidare klarade terrängmodellen producerad med Dji Phantom endast kravet för asfaltsytor, där medelavvikelsen fastställdes till 0,001 m. Terrängmodellen producerad med Smartplanes klarade endast kraven för marktyperna asfalt och grus där medelavvikelsen fastställdes till -0,007 m respektive 0,017 m. Som en del i studien jämfördes programvarorna PhotoScan och UASMaster för framställning av punktmoln för bilder insamlade med Smartplanes S1C. Resultatet visar att PhotoScan uppnådde lägst mätosäkerhet för asfalt, gräs och grus medan UASMaster uppnådde lägst mätosäkerhet för naturmark. Studien visar att flygburen laserskanning borde vara en fortsatt föredragen metod för insamling av topografisk data då metoden resulterade i lägst mätosäkerheter i denna studie. Vidare visar studien att det är möjligt att framställa digitala terrängmodeller med UAS för detaljprojektering enligt SISTS 21144:2016 för asfalt- och grusytor. Dessutom konstateras att olika bearbetningsprogram skiljer sig vid framställning av punktmoln.

Unmanned Aerial Vehicles (UAVs) är obemannade flygfarkoster som främst använts och utvecklats inom det militära. Under senare år har användandet även tagit fart inom den civila sektorn, däribland mätningsbranschen. För att samla in geodata används Unmanned Aircraft Systems (UAS), vilka är system som består av fler komponenter än endast luftfarkosten t.ex. även kamera och kontrollstation. UAS är ett bra alternativ till traditionell flygfotografering då högupplösta bilder kan genereras till en låg kostnad. Eftersom UAS är en relativt ny metod måste osäkerheten utvärderas. Syftet med detta examensarbete är att utvärdera hur stödpunkter påverkar osäkerheten vid georeferering av UAS-bilder. Data erhölls från en flygning utförd av Sweco i november 2012. För att kunna utvärdera stödpunkternas inverkan översignalerades det 5 ha stora området med 35 stödpunkter. Nio olika konfigurationer av stödpunkter georefererades i programvaran Agisoft PhotoScan 0.9.0 och resultatet analyserades i Microsoft Excel, Geo Professional och Surfer 10L. Resultaten visar att osäkerheten för georefereringen minskar när antalet stödpunkter ökar, förutsatt att en jämn placering tillämpas. Bra georeferering uppnåddes när fyra stödpunkter användes. Vi rekommenderar ändå att minst fem stödpunkter används, fem stycken ger bra möjligheter till en god geometri – en i varje hörn och en i mitten. Det lägsta RMS-värdet i 3D (72 mm) erhölls med 17 stödpunkter jämnt fördelade över området. Det högsta RMS-värdet i 3D (190 mm) fick konfigurationen med sex stödpunkter placerade i ett av områdets hörn, något som tydligt visar hur stödpunkters placering (geometrin) påverkar osäkerheten av georefereringen. Även om fyra stödpunkter (en i varsitt hörn) bara får marginellt större RMS-värde än om en extra stödpunkt placeras i mitten, rekommenderas den sistnämnda för den bättre geometrin. För att kontrollera georefereringen rekommenderas några extra inmätta kontrollpunkter i området. / The main use and development of Unmanned Aerial Vehicles (UAVs) havethrough history been driven for military purposes, but in recent years the usehas increased also in the civilian sector, including the surveying industry. Inorder to collect geodata Unmanned Aircraft Systems (UAS) are used. UAS aresystems that consist not only of the unmanned vehicle, but also of componentslike a camera and a control station. UAS is a good alternative to traditionalaerial survey due to the high resolution images and the low operational cost.The uncertainty of UAS must be evaluated further since it is a relatively newsurveying method. The purpose of this study is to analyze the number of groundcontrol point’s (GCP’s) impact on the uncertainty of georeferencing UAS images.Data was collected from a flight conducted by Sweco in November 2012. The areawhich was flown (5 ha) was “over-signalized” by 35 GCPs in order to evaluate theirimpact on the georeferencing uncertainty. Nine different configurations of GCPswere georeferenced in the software Agisoft PhotoScan 0.9.0 and the result wasanalyzed in Microsoft Excel, Geo Professional and Surfer 10L. The result showsthat the uncertainty of the georeferencing decreases when the number of GCPsincreases, provided their distribution is even in the area. A goodgeoreferencing was obtained when four GCPs were used. Regardless, we recommendthe use of five, five provide a good geometry – one in each corner and one inthe middle. The least RMS value in 3D (72 mm) was found with 17 GCPs evenlydistributed in the area. The highest RMS value in 3D (190 mm) was found whenall six GCPs were placed in one of the corners of the area. This shows that thedistribution of GCPs has a great impact on the uncertainty of thegeoreferencing. Even if four GCPs (one in each corner) just get a little higherRMS value than if one extra GCP is placed in the middle, the latter isrecommended because of the favourable geometry. To be able to control thegeoreferencing it is recommended to survey some extra GCPs in the area.

With the continued rise in wildfires in California, and around the world, technological advancements are needed to improve the safety and effectiveness of wildland firefighters. One area that provides an opportunity for such development is the deployment of temporary communications networks. Currently, radio repeaters are set up on mountain tops in the response area; such repeaters do not provide flexibility once installed, still have blind spots, and require the time of valuable assets like helicopters to install. This thesis will establish the feasibility of airborne radio repeaters for wildland firefighting. In order to successfully demonstrate the feasibility of such an airborne system, the resulting system should be rapidly deployable, improve communications range and reliability, and be compatible with existing regulations and guidelines. The design process for the repeater payload is described, as well as important troubleshooting steps. The resulting product is then compared to the initial requirements through testing and observation. Although audio filtering provided by off-the-shelf handheld radios prevented the repeater from functioning as intended, the proposed 2m/70cm dual-band digital communications relay was capable of being carried by the Altavian Nova and was able to successfully demonstrate the feasibility of such a system. As such it will be an important contribution to communications needed for fighting future wildfires.

Energy-shaping techniques are used to expand the range of autonomous motion of unmanned aerial systems without prohibitively {color{black}increasing the computational cost of the resultant controller}. Passivity-based control presents a method to implement a static, nonlinear state feedback control law that stabilizes the motion of an aircraft with a large region of attraction. {color{black} The energy-based control scheme is applied to both multirotor and fixed-wing aircraft}. Multirotor aircraft dynamics are cast into a port-Hamiltonian System and the concept of trajectory tracking using canonical feedback transformation is implemented to construct a cross-track controller. Fixed-wing aircraft dynamics are cast in port-Hamiltonian form and a passivity-based nonlinear control law for steady, wings-level flight of a fixed-wing aircraft to a specified inertial velocity (speed, course, and climb angle) is constructed. Results in simulations and experiments suggest robustness, and a large region of attraction of the controller. The control law extended to support time-varying inertial velocity tracking that incorporates banking to turn. The results are extended by including a line-of-sight guidance law and varying the direction as a function of position relative to a desired path, rather than as a function of time. The control law and the associated proof of stability follow similarly to that of the time-varying directional stabilization problem. The results are supported with simulations as well as experimental flight tests. / Doctor of Philosophy / This dissertation presents an alternative but intuitive approach to regulate unmanned aerial vehicles' flight that would allow for more maneuverability {color{black} than conventional methods}. This scheme relies on modifying the energy of the system to achieve the desired motion and leverages the properties of the aircraft rather than eliminating them and imposing different properties. This approach is applied to both fixed-wing and aircraft and quadcopters. Simulations and experimental flights have show the efficacy of this approach compared to other more established methods.