We define a novel algorithm based on utility functions for dynamically allocating tasks to mobile robots in a multi-robot system. The algorithm attempts to maximize the performance of the mobile robot while minimizing inter-robot communications. The algorithm takes into consideration the proximity of the mobile robot to the task, the priority of the task, the capability required by the task, the capabilities of the mobile robot, and the rarity of the capability within the population of mobile robots. We evaluate the proposed algorithm in a simulation study and compare it to alternative approaches, including the contract net protocol, an approach based on the knapsack problem, and random task selection. We find that our algorithm outperforms the alternatives in most metrics measured including percent of tasks complete, distance traveled per completed task, fairness of execution, number of communications, and utility achieved.

Multi-robot systems leverage the numbers and characteristics of different robots to accomplish an overall mission. Efficient task allocation and motion planning of multi-robot teams are essential to ensure each robot's actions contribute to the overall mission while avoiding conflict with each other. The original contribution of this thesis is an optimized, efficient, and multi-factor task allocation algorithm to comprise the main component of a task coordination framework (TCF), with motion planning as a secondary component. This algorithm determines which robot performs which tasks and in what order. It presents a novel solution to the multiple robot task allocation problem (MRTA) as an extension of the multiple travelling salesmen (MTSP) problem. This extension to the MTSP considers operational factors representing physical limitations, the suitability of each robot, and inter-task dependencies. The task allocation algorithm calculates an optimized distribution of tasks such that a global objective function is minimized to simultaneously reduce total cost and ensure an even distribution of tasks among the agents. Once an optimized distribution of tasks is calculated, the motion planning component calculates collision-free velocities to drive the robots to their goal poses to facilitate task execution in a shared environment. The proposed TCF was implemented on teams of unmanned air vehicles (UAVs) and unmanned ground vehicles (UGVs). Test cases considered scenarios where the UAVs executed aerial observation tasks while UGVs executed simulated patrol and delivery tasks. The solutions were tested using real-life robots as a proof of concept and to validate simulations. The robots' kinematic and computer vision models were combined with the task coordination framework to facilitate the implementation. Large-scale simulations involving greater numbers of robots operating in a larger area were also conducted to demonstrate the task coordination framework's versatility and efficacy.

No description available.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第24103号 / 工博第5025号 / 新制||工||1784(附属図書館) / 京都大学大学院工学研究科航空宇宙工学専攻 / (主査)教授 藤本 健治, 教授 太田 快人, 准教授 丸田 一郎, 教授 泉田 啓 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

To reduce complexity of system analysis and control design, simplified models that capture the behavior of interest in the original system can be obtained. These simplified models, called abstractions, can be analyzed more easily than the original complex model. Hierarchies of consistent abstractions can significantly reduce the complexity in determining the reachability properties of nonlinear systems. Such consistent hierarchies of reachability-preserving nonlinear abstractions are considered for the robotic car. Not only can these abstractions be analyzed with respect to some behavior of interest, they can also be used to transfer control design for the complex model to the simplified model. In this work, the abstraction is applied to the car/unicycle system. Working towards control design, it is seen that there are certain classes of trajectories that exist in the rolling disk system that cannot be achieved by the robotic car. In order to account for these cases, the new concepts of traceability and &#949-traceability are introduced. This work also studies the relationship between the evolution of uncertain initial conditions in abstracted control systems. It is shown that a control system abstraction can capture the time evolution of the uncertainty in the original system by an appropriate choice of control input. Abstracted control systems with stochastic initial conditions show the same behavior as systems with deterministic initial conditions. A conservation law is applied to the probability density function (pdf) requiring that the area under it be unity. Application of the conservation law results in a partial differential equation known as the Liouville equation, for which a closed form solution is known. The solution provides the time evolution of the initial pdf which can be followed by the abstracted system. / Ph. D.

This dissertation is concerned with a space robot consisting of a rigid platform, two articulated flexible arms and a rigid end-effector. The task is to ferry some payload and to dock smoothly with an orbiting target whose motion is either known or not known a priori. The dynamical equations for planar motion of the space robot are derived by means of Lagrange’s equations. They are then separated into two sets of equations suitable for rigid-body maneuver control design and vibration suppression control design. A perturbation method is used when the target motion is known a priori and direct partitioning is used when the target motion is not known. Both approaches are under the assumption that maneuver motions are much larger than elastic motions. As far as the rigid-body maneuver control is concerned, optimal trajectory planning is carried out off-line by means of the global optimization method under the assumption that the target motion is known a priori. In contrast, when the target motion is not known a priori, on-line feedback tracking control is carried out by means of an algorithm based on Liapunov-like methodology and using on-line measurements of the target motion. As far as the vibration suppression control is concerned, the use of the piezoelectric sensor/actuator pairs dispersed along the flexible arms is proposed. Collocated sensors/actuators for vibration control exhibit good performance. The actuators are designed to compensate for the disturbances caused by the rigid-body maneuver and to realize the LQR feedback control. Assuming that the number of actuators along each flexible arm is equal to the number of modes used to model the beam, the LQR control design is based on a linear time-varying system without persistent disturbances. Problems related to the digital implementation of the control algorithms are also discussed. Some undesirable effects, such as the bursting phenomenon and even system instability, can occur if the control algorithms are realized in discrete-time. To prevent these problems, the modified discrete-time control schemes are developed. Numerical examples are used to demonstrate the control algorithms. / Ph. D.

A distributed paradigm is proposed for behavior-based control of a homogeneous collection of autonomous mobile robots in the lifting and lowering processes of payload transportation. Unlike previous applications of behavior-based control to payload transportation, we examine control of a payload in a vertical plane. Others before have examined moving payloads on a horizontal surface through pushing actions; we demonstrate an ability to both raise and lower a pallet, despite the fact that no robots have a rigid grasp of the pallet. This control paradigm uses parallel behavior pathways within the individual robot and minimal emergent specialization between robots to control both pallet translation and rotation, while maintaining a strong tolerance to environmental uncertainties and changes. We stress simple, feasible methodologies over complex, optimal methodologies, although we show that with some global self-organization of the collective, the feasible solutions approach and become optimal solutions. These mobile robots demonstrate an ability to function in unforeseen environments and with inaccurate sensor data. They also demonstrate an ability to learn their place, or role, within the collective. The robots must learn their relative roles because they possess no predetermined knowledge about pallet mass, pallet inertia, collective size, or their positions relative to the pallet's center of gravity. All of this is achieved using memoryless, behavior-based control algorithms with minimal inter-agent communication. / Master of Science

The Autonomous Challenge component of the Intelligent Ground Vehicle Competition (IGVC) requires robots to autonomously navigate a complex obstacle course. The roadway-type course is bounded by solid and broken white and yellow lines. Along the course, the vehicle encounters obstacles, painted potholes, a ramp and a sand pit. The success of the robot is usually determined by the software controlling it. Johnny-5 was one of three vehicles entered in the 2004 competition by Virginia Tech. This paper presents the vision processing software created for Johnny-5. Using a single digital camera, the software must find the lines painted in the grass, and determine which direction the robot should move. The outdoor environment can make this task difficult, as the software must cope with changes in both lighting and grass appearance. The vision software on Johnny-5 starts by applying a brightest pixel threshold to reduce the image to points most likely to be part of a line. A Hough Transform is used to find the most dominant lines in the image and classify the orientation and quality of the lines. Once the lines have been extracted, the software applies a set of behavioral rules to the line information and passes a suggested heading to the obstacle avoidance software. The effectiveness of this behavior-based approach was demonstrated in many successful tests culminating with a first place finish in the Autonomous Challenge event and the $10,000 overall grand prize in the 2004 IGVC. / Master of Science

A method for coordinating a homogeneous swarm of autonomous mobile robots is presented. The broadcast-based coordination scheme was developed for the Army Ant swarm—a system of small, relatively inexpensive mobile robots that can accomplish complex tasks by cooperating as a team. The primary drawback of the Army Ant system is that the absence of a central supervisor poses difficulty in the coordination and control of the agents. Our coordination scheme provides a global "group dynamic" that controls the actions of each robot using only local interactions. Coordination of the swarm is achieved with signals we call "heartbeats". Each agent broadcasts a unique heartbeat and responds to the collective behavior of all other heartbeats. We generate heartbeats with van der Pol oscillators, which are nonlinear oscillators that modify their output when coupled to other oscillators. Van der Pol oscillators have long been utilized in simulations, particularly to model rhythmic behavior in biological systems. In this application, we use the known properties of coupled van der Pol oscillators to create predictable group behavior. We emphasize the use of this controller to allow agents to simultaneously perform an action such as lifting, steering, or changing speed. For this research we synthesize a three oscillator network to show that we can achieve multi-agent coordination. An inexpensive FM communication link is used to broadcast and receive oscillator signals. We show that the network may be configured to entrain to a leader or to a common frequency. Additionally, we use our coordination scheme to provide global speed control to our three agent system. / Master of Science

Orientadores: Paulo Augusto Valente Ferreira, Alessandro Correa Victorino, Samuel Siqueira Bueno / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-08T18:02:11Z (GMT). No. of bitstreams: 1 Castro_CesarDantasde_M.pdf: 3031758 bytes, checksum: 02179aa87aa18297b9b19bef7fb5b647 (MD5) Previous issue date: 2007 / Resumo: Este trabalho disserta sobre o desenvolvimento e a implementação de um sistema de localização e mapeamento simultâneos (SLAM) para um veículo robótico aéreo. Utilizando tal sistema, um robô que sobrevoe determinada área, até então desconhecida, deve ser capaz de conhecer sua postura no ambiente e mapeá-lo, sem o auxílio de mapas ou outras informações externas. Para alcançar este objetivo, o sistema recebe informações de uma unidade de medição inercial e de uma câmera, que observa características do ambiente e, indiretamente, a posição e a atitude do robô. Para fundir as informações dos dois conjuntos sensoriais embarcados, é utilizada uma arquitetura baseada no filtro de Kalman estendido, que atua como um estimador tanto da localização do dirigível quanto do mapa. Este sistema representa um primeiro passo em direção a uma solução de SLAM em seis graus de liberdade para o Projeto AURORA, que visa o desenvolvimento de tecnologia em robótica aérea. Desta forma, a abordagem proposta é validada em um ambiente de simulação composto de sensores virtuais e do simulador dinâmico do projeto AURORA. Os resultados apresentados mostram a eficácia da metodologia / Abstract: This work addresses the development and implementation of a simultaneous localization and mapping (SLAM) system for aerial robotic vehicles. Through this system, a robot flying over an unknown region must be capable of detecting its position accurately and, at the same time, constructing a map of the environment without the help of maps or any other external information. To reach that goal, the system receives input data from an inertial measurement unit and a single camera, which observes features in the environment and, indirectly, the robot¿s position and attitude. The data from both onboard sensors are then fused using an architecture based on an extended Kalman filter, which acts as an estimator of the robot pose and the map. This system represents a first step towards a six degrees of freedom SLAM solution for Project AURORA, whose goal is the development of technology on aerial robotics. As such, the proposed methodology is validated in a simulation environment composed of virtual sensors and the aerial platform simulator of the AURORA project based on a realistic dynamic model. The reported results show the efficiency of the approach / Mestrado / Automação / Mestre em Engenharia Elétrica