Дисертації з теми "Networked teleoperation"

La téléopération en réseau est une thématique émergente, où un humain (le maître) communique avec un esclave commandé à distance à travers un réseau de communication, pour l'échange des données de commande et de mesure. Pour la téléopération longue distance, ces informations traversent divers réseaux hétérogènes ou hybrides interconnectés ensemble. Nous traitons également le cas où un segment sera sans fil, ce qui permet d'envisager que l'esclave soit mobile (véhicule télé-opéré par exemple). Une étude indépendante des aspects commande et des aspects réseau entraîne l'atteinte d'objectifs locaux, qui peuvent le cas échéant être ensuite fortement dégradés après l'intégration des deux aspects. L'utilisation de réseaux partagés, comme internet, en lieu et place d'une liaison dédiée (point à point, liaison virtuelle) est également un défi, en particulier dans le cas de la longue distance. Les mécanismes de qualité de service (QdS) d'un réseau, en particulier du réseau hétérogène, doivent pouvoir être traités par les autres réseaux ; ces points sont assez souvent traités avec un point de vue d'applications multimedia, plus rarement avec le point de vue de la téléopération critique. L'approche de téléopération en réseau peut profiter de ces mécanismes de qualité de service en utilisant des priorités ou des réservations de bandes passantes en fonction des critères de commande et de transparence. Cela peut entraîner une adaptation des ressources réseau en fonction des besoins des objectifs de téléopération afin de transporter l'information en satisfaisant les besoins bout en bout de l'application, cette notion devant être précisée. Lorsqu'on considère la QdS avec un point de vue réseau, elle consiste principalement à la gestion de flux de communication, en prenant en compte des paramètres de besoin et d'ordonnancement. Si l'on prend en compte le point de vue de la téléopération, la QdS entraîne des conséquences sur la qualité de commande (QdC) ce qui inclut la stabilité, la transparence et la téléprésence. De plus, l'interface humaine peut également avoir des besoins spécifiques en termes de téléprésence, ce qui peut entraîner des conséquences sur la répartition de la charge du réseau. Afin d'intégrer les deux approches de QdS et de QdC pour apporter une meilleure réponse au problème posé, nous proposons une approche de coconception avec d'une part une adaptation de la QdS aux besoins de la Qdc (qualité de la transparence) et d'autre part une adaptation de la Qdc (par exemple dégradation de la qualité du flux vidéo) à la disponibilité du réseau. Nous avons considéré deux cas de figures. Dans un premier temps, sans tenir compte des mécanismes de QdS du réseau (approche dite du meilleur effort ou best effort) et par l'adaptation au mieux de l'application, pour répondre aux besoins de la téléopération. La seconde approche considère un réseau orienté QdS entre le maître et l'esclave, où différentes priorités peuvent être attribuées aux flux nécessaires à la téléopération, en fonction des contextes. L'approche proposée est finalement mise en oeuvre sur NeCS-Car, la voiture téléopérée disponible au Laboratoire
Networked teleoperation (NT) is an emerging area of technology, where human assisted Master and remote Slave devices communicate over a communication network for the exchange of command and sensor feedback information. For long range mobile teleoperation, this information travels over different types of heterogeneous/hybrid networks interconnected together with a compulsory network segment over wireless to permit increased mobility. The independent design of control and network promotes the isolated objectives and the performance is degraded after an integration. It is also challenging if internet is used for teleoperation, instead of point to point communication. The mechanisms of QoS in one network protocol of the heterogeneous network needs mapping on any other network which is usually carried out with a multimedia point of view and not for the critical teleoperation data. The approach for networked teleoperation can be given a new dimension by adding quality of service (QoS) to different flows on need based priority and as a function of control and transparence criteria. This means to alter the network resources for teleoperation objective in order to transport the information to satisfy the end-to-end application needs. The network QoS perspective is important to consider in the co-design approach for teleoperation. However, the true meaning of end to end QoS must be defined for teleoperation. If we consider the QoS from the network point of view, it refers to the management of various flows or users as per their need and precedence. Whereas, from the teleoperation perspective, QoS is related to the quality of control (QoC) which includes stability, transparence and telepresence. Moreover, the human interface may have different requirements as per telepresence feeling which will result in varying network load. To control network QoS, we need to measure or estimate it. Therefore, classification of QoS is performed with a fuzzy inference system which is able to distinguish between varying levels of QoS. In addition, it is also a challenging task for online decision for reconfiguration of network and control performance. We have used supervised methods for classification and prediction of QoS to be used in the proposed approach. Thus, data communication networks treat control information as best effort most of the time. Conclusion the communication used in teleoperation uses a dedicated network/communication. In short, a co-design approach is formulated to treat the network QoS as a function of teleoperation objectives which are related to the quality of transparence and control. Alternatively, the video flow is managed in order to effectively reduce the necessary throughput for instants when the network quality is not sufficient. We have considered two cases. First, without taking into account any network QoS mechanism (best effort flows only) and adapting application needs as per the teleoperation objectives. The second one considers a QoS oriented network in between the master and slave, where different priorities can be assigned to the teleoperation flows as per need. In the end, the proposed methodology is implemented on the NeCS-Car benchmark

A bilateral teleoperation system makes it possible for a human operator to interact with a remote environment and receive feedback from their actions. In this work, two different teleoperation architectures are studied, one well-established approach called position-force, and another approach called Model-Mediated Teleoperation (MMT). Position-force generates force feedback by measuring forces and sending them over the network, subject to network delay, while MMT generates nondelayed force feedback through a local environment model. The two architectures are compared in terms of transparency (quality of force feedback) and state consistency (equivalence of positions) under different network conditions. Up to 150 ms constant delay in Round Trip Time (RTT) is added. Both architectures are implemented as relatively simple versions of themselves. The case (teleoperation task) under consideration is pushing a cuboid object on a planar surface in a single direction. The MMT architecture is also studied in further detail, in terms of model complexity. Two versions of an MMT model are investigated, one simpler model that has a linear representation of energy loss and one less simple model that has a non-linear representation of energy loss. The purpose is to see what implications there are of increasing or decreasing model complexity. The results indicate that position-force has better performance in terms of both transparency and state consistency than both MMT models for all investigated network conditions. The simple version of MMT (linear model) performs better than the non-linear model, in terms of transparency and state consistency of cuboid positions, for all network conditions except the largest added delay. In terms of state consistency for the device positions, the non-linear model only performs better than the linear model with no added delay.
En distansstyrt system med tvåsidig kommunikation gör det möjligt för en mänsklig operatör att interragera med en avlägsen miljö och få återkoppling. I detta arbete studeras två olika arkitekturer för dessa typer av system. Ett väletablerat tillvägagångssätt som kallas position-force, och ett annat som kallas Model-Mediated Teleoperation (MMT). Position-force genererar kraftåterkoppling genom att mäta krafter och skicka dem över nätverket, utsatt för nätverksfördröjningar, medan MMT genererar en icke-fördröjd kraftåterkoppling från en lokalt modellerad miljö. De två arkitekturerna jämförs med avseende på ”transparency” (kvalitén på kraftåterkopplingen) och ”state consistency” (matchning av positioner) under olika nätverksförhållanden. De studerade förhållandena innefattar upp till 150 ms konstant fördröjning i tur och returtid. Båda arkitekturer implementeras som förhållandevis enkla versioner av sig själva. Det studerade fallet (operatörens uppgift) innefattar puttande på ett rätblock på en plan yta i en enkel riktning. MMT arkitekturen studeras även i vidare detalj med avseende på komplexitet hos den lokala modellen. Två versioner av MMT modellen undersöks, en enklare model som har en linjär representation av energiförlust, och en mer avancerad model som har en icke-linjär representation av energiförlust. Avsikten bakom detta är undersöka vilka implikationer som finns när komplexiteten hos modellen höjs eller sänks. Resultaten indikerar att position-force presterar bättre än MMT med avseende på både transparency och state consistency för alla undersökta nätverksfördröjningar. Den linjära MMT modellen presterar bättre än den icke-linjära modellen, med avseende på transparency och state consistency i rätblockspositioner, för alla nätverksförhållnaden utom den största pålagda fördröjningen. Vidare presterar den icke-linjära modellen bättre än den linjära modellen, med avseende på state consistency för styrande/styrd apparat, endast då ingen nätverksfördröjning läggs på.

With the continuous progress on autonomous vehicle and remote drivingtechniques, connection quality demands are changing compared withconventional quality of service. Vehicle to everything communication, asthe connectivity basis for these applications, has been built up on LongTerm Evolution basis, but due to various ethical and environmental issues,few implementations have been made in reality. Therefore simulation approachesare believed to provide valuable insights.To fully model an LTE vehicular network, in this work we first providea comparison study to select the preferable LTE simulator. Aimingto integrate communication nodes with mobility, a solution for simulationframework is developed based on a state-of-art comparison study on theexisting simulator frameworks. We then further develop the network simulator,and complement it with hybrid wireless channel modeling, channeland quality of service aware scheduler, and admission control strategies. Interms of instant optimization of the network, real-time access is emulatedfor external devices to communicate with the simulator. In this thesis,the evaluation of the framework performance considers two aspects: theperformance of the simulator in LTE V2X use case and the feasibility ofthe service, specifically, remote driving, under realistic network capacity.For our framework, the results indicate that it is feasible to realize remotedriving in an LTE urban scenario, but, as an example, we show that foran area of Kista, five vehicles could be hold by a base-station with guaranteedservice at most.
Med kontinuerliga framstegen p°a autonomt fordon och fj¨arrkontrollteknikf¨or¨andras kravet p°a anslutningskvalitet i j¨amf¨orelse med konventionell servicekvalitet.Fordon till allting (V2X) kommunikation, som anslutningsgrundf¨or dessa applikationer, har byggts upp p°a basis av Long TermEvolution (LTE) system, men p°a grund av olika etiska och milj¨om¨assigaproblem har f°a implementeringar gjorts i verkligheten. D¨arf¨or antas simuleringsmetoderge v¨ardefulla insikter.Att fullt ut modellera ett LTE-fordon n¨atverk, i det h¨ar arbetet ger vif¨orst en j¨amf¨orelsestudie f¨or att v¨alja den f¨oredragna LTE-simulatorn.I syfte att integrera kommunikationsnoder med r¨orlighet utvecklas enl¨osning f¨or ett simuleringsramverk baserat p°a en j¨amf¨orelsestudie p°a befintligasimulatorramar. Vi utvecklar sedan n¨atverkssimulatorn ytterligare,och kompletterar den med hybrid tr°adl¨os kanalmodellering, kanal ochservicekvalitetmedvetna schemal¨aggning och antagningskontrollstrategier.N¨ar det g¨aller direkt n¨atverksoptimering, emuleras realtidsanslutningav externa enheter f¨or att kommunicera med simulatorn. I denna avhandlingutv¨arderas ramverken i tv°a aspekter: simulatorns prestanda i LTEV2X-anv¨andningsomr°adet och genomf¨orbarheten av tj¨ansten, s¨arskilt fj¨arrk¨orning,under realistisk n¨atkapacitet. In v°ara ramverk visar resultaten att det ¨arm¨ojligt att realisera fj¨arrk¨orning i ett LTE-urbana scenario, men som exempelvisar vi att f¨or ett omr°ade i Kista skulle som mest fem fordon kunnask¨otas av en basstation med garanterad service.

Teleoperation system is an important tool to control a device or model in an isolated area remotely where the operator cannot perform the task locally. The vast majority of teleoperation systems provides the operator with visual and haptic control to accomplish the assignment as naturally as possible. However, on a teleoperation system with considerable distance, the time delay could cause a drop in performance. This thesis aims to minimize delay problem by implementing a prediction approach using Kalman Filter. Kalman Filter algorithm has been widely used to estimate user movement for tracking systems. Kalman filter provides an efficient mechanism to predict future state based on Bayesian estimation to sequentially predict future states and measure an actual system to update system parameters. The primary objective of this work is to extract information generated by our prototyping model and visualizing the data to reflect the performance of the system. We use Phantom Omni devices and 3D arm as a model. Different type of Kalman filter algorithms is used to test the accuracy and performance of predicted state generated by the filter. The result shows that the implementation of Extended Kalman Filter (EKF) and smoothing function could overcome the networking delay on certain degrees. The comparison shows that the EKF has better accuracy and performance compared to Unscented Kalman Filter (UKF) when estimating the future state. Additionally, the implementation of smoothing function could improve the stability of teleoperation system.
Teleoperation systemet är ett viktigt verktyg för att styra en enhet eller modell i ett isolerat område på distans där operatören inte kan utföra uppgiften lokalt. De allra flesta av teleoperation system ger föraren visuell och haptisk kontroll för att utföra uppdraget så naturligt som möjligt. Men på en teleoperation system med stort avstånd, kan tidsfördröjningen medföra en nedgång i prestanda. Denna avhandling syftar till att minimera förseningar problem genom att implementera en förutsägelse tillvägagångssätt med Kalman Filter. Kalman filteralgoritm har i stor utsträckning används för att uppskatta användarens rörlighet för spårning. Kalman filter ger en effektiv mekanism för att förutsäga framtida stat grundad på Bayesian uppskattningen att sekventiellt förutsäga framtida tillstånd och mäta ett verkligt system för att uppdatera systemparametrar. Det primära syftet med detta arbete är att extrahera information som genereras av vår prototypmodell och visualisera data för att återspegla systemets prestanda. Vi använder Phantom Omni enheter och 3D-arm som en modell. Olika typer av Kalman filter algoritmer används för att testa riktigheten och prestandan hos förutsagda tillståndet genereras av filtret. Resultatet visar att genomförandet av Extended Kalman filter (EKF) och utjämningsfunktionen kan övervinna nätverk dröjsmålsvissa grader. Jämförelsen visar att EKF har bättre noggrannhet och prestanda jämfört med Unscented Kalman Filter (UKF) vid bedömningen av framtida tillstånd. Dessutom, genomförandet av utjämningsfunktionen skulle kunna förbättra stabiliteten hos teleoperation systemet.

An interesting approach in cooperative control is to design distributed control strategies which use only local information so that a multi-agent system achieves specified behaviors. A basic behavior in cooperative control is the consensus. Given a multi-agent system, like a multiple robot network, it is said that the agents reach a consensus if the state of each agent converges to a common state. Examples of cooperative tasks in which consensus algorithms are employed include formation control, flocking theory, rendezvous problems and synchronization. These cooperative tasks have several possible applications, like: transportation systems (intelligent highways, air-traffic control); military systems (formation flight, surveillance, reconnaissance, cooperative attack and rendezvous) and mobile sensor networks (space-based interferometers, environmental sampling). The solution to the consensus problems involves the design of control algorithms such that the agents can reach an agreement on their states. There are two main problems that are studied in consensus, the leader-follower consensus and the leaderless consensus. In the leader-follower consensus problem, there exists a leader that specifies the state for the whole group while in a leaderless consensus problem, there is not a priori reference state. The main goal of this thesis is the design of operational space controllers that solve the leader-follower and the leaderless consensus problems in networks composed of multiple heterogeneous robots. Furthermore, this document proposes novel operational space control schemes for bilateral teleoperation systems. In both scenarios, different conditions are studied, such as the absence of robot velocity measurements, constant and variable time-delays in the robot's interconnection, and uncertainty in the robot's physical parameters. Most of the previous consensus control algorithms, only work with the position or orientation but not with both. On the contrary, this dissertation deals with the entire pose of the robots that contains both the position and the orientation. Moreover, in order to render a singularity-free description of the orientation, the unit-quaternions are employed. The dissertation provides a rigorous stability analysis of the control algorithms and presents simulations and experiments that validate the effectiveness of the proposed controllers.
Un enfoque interesante en el control cooperativo es el diseño de estrategias de control distribuido que requieran sólo información local para que un sistema multi-agente logre comportamientos específicos. Un comportamiento básico del control cooperativo es el consenso. Dado un sistema multi-agente, como una red de múltiples robots, se dice que los agentes llegan a un consenso si el estado de cada agente converge a un estado común. Algunos ejemplos de tareas cooperativas en las que los algoritmos de consenso son utilizados son los siguientes: el control de la formación, flocking, rendezvous y sincronización. Estas tareas cooperativas tienen varias aplicaciones posibles, como: sistemas de transporte (carreteras inteligentes , control de tráfico aéreo); sistemas militares (vuelo en formación, vigilancia, reconocimiento, ataque cooperativo) y redes de sensores móviles (interferómetros en el espacio, el muestreo del ambiente). La solución a los problemas de consenso implica el diseño de algoritmos de control de tal manera que los agentes pueden llegar a un acuerdo sobre sus estados. Hay dos problemas principales que se estudian en el consenso, el consenso líder-seguidor y el consenso sin líder. En el problema de consenso líder-seguidor, existe un líder que especifica el estado de todo el grupo, mientras que en un problema de consenso sin líder, no hay ningún estado de referencia definido a priori. El objetivo principal de esta tesis es el diseño de controladores en el espacio operacional que resuelvan los problemas de consenso líder-seguidor y sin líder en redes compuestas de múltiples robots heterogéneos. Además, este documento propone novedosos esquemas de control en el espacio operacional para sistemas de teleoperación bilateral. En ambos escenarios, se estudian diferentes condiciones, tales como la ausencia de medidas de velocidad de los robots, retardos constantes y variables en la interconexión de los robots y la incertidumbre en los parámetros físicos de los robots. La mayoría de los anteriores algoritmos de control que resuelven el consenso, sólo trabajan con la posición o la orientación, pero no con ambos. Por el contrario, esta tesis doctoral se ocupa de toda la pose de los robots que contiene tanto la posición y la orientación. Por otra parte, a fin de usar una representación de la orientación libre de singularidades, se emplean los cuaterniones unitarios. Esta tesis doctoral proporciona un análisis riguroso de la estabilidad de los algoritmos de control y presenta simulaciones y experimentos que validan la eficacia de los controladores propuestos

Robot and sensor networks are needed for safety, security, and rescue applicationssuch as port security and reconnaissance during a disaster. These applications rely on realtimetransmission of images, which generally saturate the available wireless networkinfrastructure. Knowledge-based Compression is a strategy for reducing the video frametransmission rate between robots or sensors and remote operators. Because images mayneed to be archived as evidence and/or distributed to multiple applications with differentpost processing needs, lossy compression schemes, such as MPEG, H.26x, etc., are notacceptable. This work proposes a lossless video server system consisting of three classesof filters (redundancy, task, and priority) which use different levels of knowledge (localsensed environment, human factors associated with a local task, and relative globalpriority of a task) at the application layer of the network. It demonstrates the redundancyand task filters for realistic robot search scenarios. The redundancy filter is shown toreduce the overall transmission bandwidth by 24.07% to 33.42%, and when combinedwith the task filter, reduces overall transmission bandwidth by 59.08% to 67.83%. Byitself, the task filter has the capability to reduce transmission bandwidth by 32.95% to33.78%. While Knowledge-based Compression generally does not reach the same levels ofreduction as MPEG, there are instances where the system outperforms MPEG encoding.

This work is to investigate the performance of two Kalman Filter Algorithms, namely Linear Kalman Filter and Extended Kalman Filter on control-based human motion prediction in a real-time teleoperation. The Kalman Filter Algorithm has been widely used in research areas of motion tracking and GPS-navigation. However, the potential of human motion prediction by utilizing this algorithm is rarely being mentioned. Combine with the known issue - the delay issue in today’s teleoperation services, the author decided to build a prototype of simple teleoperation model based on the Kalman Filter Algorithm with the aim of eliminated the unsynchronization between the user’s inputs and the visual frames, where all the data were transferred over the network. In the first part of the thesis, two types of Kalman Filter Algorithm are applied on the prototype to predict the movement of the robotic arm based on the user’s motion applied on a Haptic Device. The comparisons in performance among the Kalman Filters have also been focused. In the second part, the thesis focuses on optimizing the motion prediction which based on the results of Kalman filtering by using the smoothing algorithm. The last part of the thesis examines the limitation of the prototype, such as how much the delays are accepted and how fast the movement speed of the Phantom Haptic can be, to still be able to obtain reasonable predations with acceptable error rate.   The results show that the Extended Kalman Filter has achieved more advantages in motion prediction than the Linear Kalman Filter during the experiments. The unsynchronization issue has been effectively improved by applying the Kalman Filter Algorithm on both state and measurement models when the latency is set to below 200 milliseconds. The additional smoothing algorithm further increases the accuracy. More important, it also solves shaking issue on the visual frames on robotic arm which is caused by the wavy property of the Kalman Filter Algorithm. Furthermore, the optimization method effectively synchronizes the timing when robotic arm touches the interactable object in the prediction.   The method which is utilized in this research can be a good reference for the future researches in control-based human motion tracking and prediction.
Detta arbete fokuserar på att undersöka prestandan hos två Kalman Filter Algoritmer, nämligen Linear Kalman Filter och Extended Kalman Filter som används i realtids uppskattningar av kontrollbaserad mänsklig rörelse i teleoperationen. Dessa Kalman Filter Algoritmer har används i stor utsträckning forskningsområden i rörelsespårning och GPS-navigering. Emellertid är potentialen i uppskattning av mänsklig rörelse genom att utnyttja denna algoritm sällan nämnas. Genom att kombinera med det kända problemet – fördröjningsproblem i dagens teleoperation tjänster beslutar författaren att bygga en prototyp av en enkel teleoperation modell vilket är baserad på Kalman Filter algoritmen i syftet att eliminera icke-synkronisering mellan användarens inmatningssignaler och visuella information, där alla data överfördes via nätverket. I den första delen av avhandlingen appliceras både Kalman Filter Algoritmer på prototypen för att uppskatta rörelsen av robotarmen baserat på användarens rörelse som anbringas på en haptik enhet. Jämförelserna i prestandan bland de Kalman Filter Algoritmerna har också fokuserats. I den andra delen fokuserar avhandlingen på att optimera uppskattningar av rörelsen som baserat på resultaten av Kalman-filtrering med hjälp av en utjämningsalgoritm. Den sista delen av avhandlingen undersökes begräsning av prototypen, som till exempel hur mycket fördröjningar accepteras och hur snabbt den haptik enheten kan vara, för att kunna erhålla skäliga uppskattningar med acceptabel felfrekvens.   Resultaten visar att den Extended Kalman Filter har bättre prestandan i rörelse uppskattningarna än den Linear Kalman Filter under experimenten. Det icke-synkroniseringsproblemet har förbättrats genom att tillämpa de Kalman Filter Algoritmerna på både statliga och värderingsmodeller när latensen är inställd på under 200 millisekunder. Den extra utjämningsalgoritmen ökar ytterligare noggrannheten. Denna algoritm löser också det skakande problem hos de visuella bilder på robotarmen som orsakas av den vågiga egenskapen hos Kalman Filter Algoritmen. Dessutom effektivt synkroniserar den optimeringsmetoden tidpunkten när robotarmen berör objekten i uppskattningarna.   Den metod som används i denna forskning kan vara en god referens för framtida undersökningar i kontrollbaserad rörelse- spåning och uppskattning.

A big challenge regarding the development of autonomous vehicles is how to handle complex situations. If an autonomous vehicle ends up in a situation where it cannot make a decision on its own it will cause the car to stop, unable to continue driving. For these situations, human intervention is required. By making it possible to control the car remotely there is no need for an actual human in the car. Instead, a human operator can remotely control one or several cars from a distance. The purpose of this project was to identify such complex situations, evaluate remote control options and implement one of these controllers to drive the SVEA cars in the Smart Mobility Lab. After evaluation of possible remote control options, the PlayStation controller was chosen to be the simplest and most intuitive steering option. The controller was successfully implemented first in simulation and then on the SVEA cars in the Smart Mobility Lab. A test track was designed to measure the performance of the implemented controller and to be able to measure user-friendliness through a survey. It was concluded that a majority of the participants would not feel comfortable steering a real car using the PlayStation controller. However, a more extensive evaluation would be required to draw any major conclusions.

Neste trabalho é proposta uma plataforma robótica móvel, concebida de maneira modular e hierárquica, visando o estudo de diversos aspectos aplicados à navegação, tanto autônoma quanto semi-autônoma, em ambientes internos. O sistema proposto possibilita a implementação de arquiteturas reativas e híbridas com aprendizagem, sendo a importância e limitações desta última discutidas. Utilizando a plataforma desenvolvida, uma aplicação de navegação robótica com aprendizagem supervisionada é realizada, usando sensores de ultra-som e através de tele-operação. O objetivo é fazer com que o agente associe, em tempo real, suas próprias respostas sensoriais com as ações motoras realizadas pelo tele-operador, permitindo que a tarefa seja repetida autonomamente com alguma generalização. Para realizar tal mapeamento, uma rede de função de base radial (RBF), usando um algoritmo de aprendizado seqüencial, é apresentada e utilizada.
This work presents a mobile robotic platform, built as a modular and hierarchical approach, aiming at the study of several aspects of indoor navigation. The proposed system allows the implementation of reactive and hybrid architectures with learning, for autonomous or semi-autonomous navigation. The importance and limitations of the learning characteristics are discussed. An application of robotic navigation with supervised learning is implemented using ultrasonic sensors and teleoperation. The aim is the agent to associate, in real time, its own sensorial perception to the motor actions realized by a teleoperator, allowing the task to be repeated in an autonomous way, with some generalization. To make the corresponding mapping, a radial basis function network (RBF), trained by a sequential learning algorithm, is presented and used.

碩士
國立成功大學
機械工程學系碩博士班
92
Despite rapid advances in life-support technology, there are still many environments that may be hazardous or inaccessible to humans, such as dangerous rooms in a nuclear plant, areas near explosives, outer space, and under water. Also the application of fully-autonomous robotic systems in the same environments remains difficult or impractical in spite of advances in robot technology. Telerobotic systems, by which a human operator can enslave a remote robot manipulator using a hand controller according to information from feedback sensors, are a practical way to combine robot technology with human versatility. Such a system can bring a human operator’s perception, judgement, and dexterity to a task while allowing the operator to control the slave manipulator from a position of safety and convenience. 　　The objective of this thesis is to advance the bilateral control method of the VR-based telerobotic system operated through the Internet. The telerobotic system adopted in this study is mainly composed of a hand controller in local site, and a robotic binocular head and an industrial robot manipulator in remote site. A six-axis force-reflecting hand controller is used to make the operator feel the contact force in remote site and enhance the manipulation with the implementation of the proposed intelligent impedance control law. A virtual scene of the remote site is created in local site by the visual information provided by the binocular head in remote site and virtual reality (VR) technique to support the operator with visual information. One image-based look-and-move control strategy is proposed to enable the robotic head to saccade and fixate the target during the teleoperation. The pose of the target is then estimated by applying static stereo analysis to determine the three-dimensional information by a process called triangulation. The VR scene is created according to the pose information of the target. 　　Finally, the theoretical results are experimentally verified on the adopted telerobotic system. Specific experimental demonstrations include the unconstrained teleoperation and the peg-in-hole task to examine the performance of the system.

碩士
國立成功大學
機械工程學系碩博士班
91
Despite rapid advances in science and technology, there are still environments that may be hazardous or inaccessible to humans, such as dangerous rooms in a nuclear plant and areas near explosives. Also, robotic systems in the same environments are unable to perform tasks autonomously, although there are great achievements in robot technology. The dexterous telerobotic systems, by which a human operator can enslave a remote mechanical hand using a hand master according to information from feedback sensors, are a practical way to combine robot technology with human versatility. Such a system can bring a human operator’s perception, judgment and dexterity to a task involving precise manipulation of delicate objects, while allowing the operator to control the mechanical hand from a position of safety and convenience. The objective of this thesis is to advance the bilateral control method of the VR-based dexterous telerobotic system operated through the Internet. A dexterous telerobotic system is first constructed in this thesis. The developed dexterous telerobotic system is mainly composed of a hand master and a virtual environment in local site, and a multifingered mechanical hand and a vision system in remote site. A virtual scene of the remote site is created in local site by visual information provided by the vision system in remote site and Virtual Reality (VR) technique, to support the operator with visual information and reduce the amount of transmitted data. A fuzzy sliding mode control algorithm is then proposed for the force and position control of the system due to the model uncertainties of the tendon-driven system. Using the hand master, the mechanical hand can be controlled to follow the operator’s hand motion by using the proposed position controller. When the mechanical hand contacts with the environment, the contact force is fed back to the operator by using the proposed force controller. Finally, the theoretical derivations for the position and force controllers based on the proposed fuzzy sliding algorithm are experimentally verified.

This thesis presents a general simulation framework used for evaluating the performance of bilateral teleoperation systems under consistent and controllable network conditions. A teleoperation system is where an operator uses a master device to control a slave robot through a communication link. The communication link between the master and slave has an important impact on the system performance. Network emulation using ns-2 has been proposed as a way of simulating the communication link. It allows for the network conditions to be controlled and for repeatable results. The proposed setup was used to test the performance of a hydraulic actuator under various conditions of wired and wireless networks. Three control schemes were evaluated using various combinations of time delay and packet loss. The system was also tested simulating wireless communication between the master and slave to determine the effects of transmission power and distance on the performance of the system.

碩士
國立成功大學
機械工程學系
103
SUMMARY This thesis proposes a novel bilateral teleoperation system by using only one controller that is collocated with the local robot. Without a controller, the position/velocity signals of the remote robot are transmitted to the central controller over a communication network. In this thesis, the central controller is accomplished by using P- and PD-controllers. By utilizing Lyapunov stability analysis, the sufficient condition for ensuring energy dissipation is obtained so that stability and position tracking are guaranteed when the communication network is subject to constant delays. Since it is possible that time delays change with network status, the stability condition for P-controller with time-varying delays is also studied in this thesis. Moreover, force reflection from the remote environment to the human operator via the local robot is proven to be equal to the environmental force when the teleoperation system is in steady state. The proposed control systems for bilateral teleoperators with single control are validated via numerical simulations and experiments. In order to obtain external forces from the human operator and environment, a simple and low-cost force sensor for PHANToM Omni haptic device is developed in this thesis. After calibration by using simple linear regression, the three-axis force sensor is utilities to record the force information in the experiment.

碩士
國立成功大學
機械工程學系
104
In this thesis, control structures are proposed to deal with network uncertainties such as packet loss, constant delays, and network congestion in bilateral teleoperation. In bilateral teleoperation, the teleoperators transmit their states or outputs to each other via network. Network uncertainties are induced from transmission distance, equipment limitation, and especially overloading network. The topic can be broken into four parts. The first part is teleoperation under constant delays and packet loss using wave-variable transformation (WVT) and wave-variable modulation (WVM). The second is teleoperation using event-triggered communication and WVT under constant delays. For the third and fourth parts, we utilize event-triggered communication on P- or PD-based controllers with damping injection under constant delays. Besides, force-feedback is acheivable with the use of the proposed P- and PD-based control architectures. The proposed teleoperation structures are analyzed using nonlinear control theory, and verified using simulations and experiments. The results are further compared and discussed. The simulation and experiment results show the asymptotical stability of the proposed structures, and external force was measured and recorded by force sensors to verify the force-feedback efficacy. Besides, the event-triggered communication lowers the transmission frequency.

碩士
國立交通大學
電控工程研究所
100
This thesis develops a control scheme composed by a model predictive controller (MPC), a communication disturbance observer (CDOB), and a feedback weighting mechanism in order to overcome the inevitable delay time in an Internet-based teleoperation system. The CDOB mainly compensates for the time-delay effect by treating it as a disturbance without the knowledge of the delay-time model. Moreover, this thesis proposes an improved scheme of CDOB in order to eliminate the steady-state error existed in a networked control system without a sufficient sampling rate. Meanwhile, the MPC adopted in this thesis predicts the best control signal in order to optimize the system performance. Generally, the time-delay effect usually decreases the transparency of the network controlled system and three assist strategies are presented in this thesis as: (1) virtual operation point, (2) prediction operation point, and (3) fine adjustment switching mechanism. Contouring motion experiments on a 2-DOF virtual manipulator validates the proposed assist mechanism.

Renewable energy technologies (RETs) such as wind and photovoltaic (PV) have great potential for reducing the cost and environmental impact of electricity production in stand-alone power systems (mini-grids). However, due to fluctuating nature of RETs, and the various constraints of diesel power plants, a suitable communication system is required for better coordinating the operation of the various sources and controllable loads so as to increase diesel fuel displacement without compromising the power quality. A framework for designing and dimensioning a communications network for supervision and control of hybrid mini-grids as well as for choosing the network’s components and interfaces (including the data acquisition systems, modems, network hubs and switches, communications media and channels, etc.) is presented. In this framework, standard technologies and readily available communications units and subsystems are chosen with a view to satisfy the requirements and assumptions. In doing so, the hardware components needed for implementing such a network are identified, and various software pieces are written to integrate all elements of the network. A set of performance measures that are normally used to assess the performance of data communications networks is identified, and their values are obtained via extensive tests.

The inverse kinematic problem is crucial for robotics. In this paper, a solution algorithm is presented using artificial intelligence to improve the pseudo-inverse Jacobian calculation for the 7-DOF Whole Arm Manipulator (WAM) and 6-DOF Titan II teleoperation system. An investigation of the inverse kinematics based on fuzzy logic and artificial neural networks for the teleoperation system was undertaken. Various methods such as Adaptive Neural-Fuzzy Inference System (ANFIS), Genetic Algorithms (GA), Multilayer Perceptrons (MLP) Feedforward Networks, Radial Basis Function Networks (RBF) and Generalized Regression Neural Networks (GRNN) were tested and simulated using MATLAB. Each method for identification of the pseudo-inverse problem was tested, and the best method was selected from the simulation results and the error analysis.From the results, the Multilayer Perceptrons with Levenberg-Marquardt (MLP-LM) method had the smallest error and the fastest computation among the other methods. For the WAM-Titan II teleoperation system, the new inverse kinematics calculations for the Titan II were simulated and analyzed using MATLAB. Finally, extensive C code for the alternative algorithm was developed, and the inverse kinematics based on the artificial neural network with LM method is implemented in the real system. The maximum error of Cartesian position was 1.3 inches, and from several trajectories, 75 % of time implementation was achieved compared to the conventional method. Because fast performance of a real time system in the teleoperation is vital, these results show that the new inverse kinematics method based on the MLP-LM is very successful with the acceptable error.

Teleoperation of hydraulic manipulators is of potential when the presence of the operator, in a remote location, is inconvenient or dangerous. Augmenting such teleoperated systems using haptic sensation will further enhance performance, safety, and convenience. The advantage of using haptic force becomes more evident when it is employed to compensate for undesirable phenomena such as existence of a delayed and lossy communication channel or excitation of the manipulator base. The focus of this thesis is on haptic-enabled control of base-excited hydraulic manipulators that are controlled through a wireless communication channel. The targeted application is live transmission line maintenance. Both unilateral and bilateral controls of teleoperated hydraulic manipulators are studied. On the unilateral front, position error is shown to be an important issue, especially when the position accuracy of the slave manipulator is violated due to fast motion of the operator’s hand at the master site, lack of responsiveness in actuation system, or poor quality of communication channel. With respect to bilateral control, three main challenges are identified, and solutions to these challenges are investigated: (i) accurate control of the slave manipulator when the communication channel is delayed and/or lossy, (ii) control of the teleoperated system when the slave manipulator is mounted atop a moving platform, and (iii) transparent force feedback to improve the position tracking of the system. First, effects of network quality and slave manipulator base excitation are examined on performance of the teleoperated system. The position error between the haptic device implement and the hydraulic manipulator end-effector is shown to increase when the network is delayed and lossy. Next, excitation of the slave manipulator base deviates the end-effector from its reference trajectory, and the position error therefore becomes larger. To alleviate the position inaccuracy, a position referenced force feedback scheme is proposed. The scheme makes the input dynamics a better match with the slave dynamics. Combined with the virtual fixture force, the virtual fixture is shown to aid the operator in following a predefined virtual fixture trajectory. Due to complexity of dynamics, performance evaluations are mostly conducted using experimental validations on actual system in a laboratory setting.

Este projecto aborda a operação e controlo de equipamentos industriais à distância e integra-se num projecto mais vasto de Sistemas de Produção Virtuais/Sistemas de Produção Distribuídos/One-Product-Integrated-Manufacturing (SPV/SPD/OPIM). O primeiro objectivo do presente projecto consistiu na análise do estado da arte da área em que se insere, nomeadamente do teletrabalho, da teleoperação de equipamentos industriais, de sensorização remota e respectivo enquadramento no projecto de sistemas SPV/SPD/OPIM. O segundo objectivo consistiu na especificação, construção e validação de uma instalação protótipo de demonstração da operação e controlo remoto aplicados a um equipamento industrial. Foi realizada a análise ao estado da arte dos conceitos e implementações relacionadas com a teleoperação, incluindo uma pequena resenha histórica dos mesmos. O enquadramento nos sistemas SPV/SPD/OPIM foi realizado, constituindo a envolvente e motivações para o desenvolvimento do presente projecto. Finalmente, a instalação protótipo foi concebida e implementada sobre a rede TCP/IP da Universidade do Minho e sobre uma fresadora de controlo numérico. Foram testados diversos elementos do sistema de sensorização remota, da interface ao operador remoto e das comunicações via TCP/IP entre a sala de operações e a sala controlo. Em conclusão, queremos realçar que é nossa convicção que as tecnologias aplicadas na elaboração da instalação protótipo constituirão estruturas base para a implementação de Sistemas de Produção Avançados que alterarão significativamente a forma como as pessoas realizam trabalho e se faz produção em ambientes industriais.
This project deals with the remote operation and control of industrial equipment within the general framework of Virtual Production Systems/Distributed Production Systems/One-Product-Integrated-Manufacturing (VPS/DPS/OPIM). The first objective of this project was the analysis of the state of the art of the subject area, namely teleworking, teleoperation of industrial equipment, remote sensoring and the respective positioning within the VPS/DPS/OPIM system’s project. The second objective was the specification, construction and validation of one prototype installation for the purpose of demonstrating the remote control and operation when applied to an industrial equipment. An analysis of the state of the art, of the concepts and implementations related to teleoperation was made, which also included a brief historical review. The positioning of this project within the framework of the VPS/DPS/OPIM system was undertaken. This constituted the context and the motivation for the development of the present project. Finally, the prototype installation was projected and implemented through the TCP/IP computer network of University of Minho and through a numerically controlled milling machine. Several tests were made to the elements of the remote sensory system, to the interface of the remote operator and TCP/IP communications between the operation and the control sites. In conclusion, we would like to express our conviction that the technologies applied in the construction of the prototype installation will be the structural basis for the implementation of Advanced Manufacturing Systems that will shape the way people work and the way manufacturing is done in industrial environments.