Th. doct.-ing.--Besançon, E.N.S.M.M., 1983. N°: 4.

Th. doct.-ing.--Besançon, 1981. N°: 115.

This thesis focuses on mechanical design and simulation studies of a quadruped robot motion control system, targeting at designing an autonomous legged robot. The designed quadruped robot with ``X"-configuration is developed for traversing rocky and sloped terrain with a static walking gait. The mechanical design of the quadruped robot is illustrated in Chapter 2, including the main body design, leg design and component selection. In the design process, appropriate mechanical structures are utilized to minimize the energy consumption. To improve energy efficiency, a set of principles is proposed. Corresponding implementations are also concretely introduced in this chapter. In addition, to simplify the mechanical structure of the quadruped robot, the mass is symmetrically distributed about the frontal and lateral planes. To improve the leg agility, the leg mass is minimized. On the one hand, the lightweight design is implemented by optimizing the mass distribution of the leg mechanism. On the other hand, the key components are assembled in the body part instead of the legs as many as possible. A sufficient leg length is also selected not only to allow the robot to step on or over obstacles, but also to avoid the leg getting caught by objects. Particularly, the leg structure is demonstrated, including the hip joint, thigh part, knee joint and limb part with a telescoping joint. When the robot sustains extensive payload, the deformed shape in joints may lead to structural failures, thereby influencing the quadrupedal locomotion. Finite element analysis (FEA) is performed when designing the structural components in reasonable structures. The design processes of the shoulder part and brass rod are demonstrated as examples. Based on the setup of loads and fixtures, the maximum deformed shape of these structural components are analyzed. From FEA simulation results, the yield strength is two orders of magnitude larger than the maximum of von Mises stress. Hence, these components are suitable to be incorporated in the quadruped robot. Based on the designed mechanical structure, simulation studies of the quadruped robot motion control system are analyzed in Chapter 3, including the modeling for a robotic leg and animated simulation. Since the quadrupedal locomotion is executed by manipulating the postures of four legs, the leg model is significant to the motion control system, thereby being analyzed mathematically. Two links kinematic conversion is implemented between the foot-end trajectory and joint angles. The dynamic model of the leg is also computed to discovery the relationship between the actuating torques and joint angles. To animate the quadrupedal locomotion, a CAD robot model is converted into MATLAB. Following the predefined footfall pattern, four legs move in sequence to execute the creeping gait. The segment of the desired trajectory of the foot-end fits a fifth order polynomial and does not include the set of singular configurations. Then, the PD control is utilized to adjust the leg posture to track the desired path. Furthermore, the actual joint angles are calculated in the MATLAB/SimMechanics quadruped robot by using Euler-Lagrange equations. Lastly, simulation results are presented to analyze the tracking performance in the joint angles and foot-ends. Finally, conclusions of the thesis are summarized, and future work is presented in Chapter 4. / Graduate / 2019-03-07

M.Ing. / This thesis deals with the modelling and control of an autonomous four-wheeled car-like vehicle. The autonomous vehicle is built on the chassis of a Tamiya TXT-1, one-tenth scale, off-road remote control truck. The simplified kinematic and dynamic models of this vehicle are derived for the purposes of control. These models are converted to a suitable form and control laws are then developed to perform set-point control as well as path following. A sonar array is constructed to permit the vehicle to sense its surroundings and construct a map of an unknown environment as it navigates through it. Lastly the Vector Field Histogram (VFH) obstacle avoidance algorithm is implemented and used to guide the vehicle through a sparsely populated environment without colliding with the obstacles.

To help control the high cost of health care delivery, a robotic system is proposed for use in passing surgical instruments in an operating room. The system consists of a vision system, a robotic arm, a speech recognition and synthesis unit, and a microcomputer. A complete vision system has been developed using standard and new techniques to recognize arthroscopic surgical instruments. Results of the vision system software evaluation gave an overall recognition accuracy of over 99%. Also, error conditions were analysed and found to be consistent with the results of a clinical survey on the proposed instrument-passing robot. As well, a payback and cost benefit analysis using estimated system costs and potential labour savings showed that the instrument-passing robot is economically feasible. Based on the results of this thesis, it was concluded that the instrument-passing robot would be beneficial for reducing the high cost of health care. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

No description available.

No description available.

Les robots parallèles entraînés par câbles sont un type de manipulateur bénéficiant d’un grand espace de travail en translation tout en offrant des performances dynamiques supérieures à celles des architectures de robots traditionnelles. Malgré tout, ceux-ci demeurent tout de même peu utilisés et les rares prototypes qui existent sont la plupart du temps trouvés en milieu académique. Plusieurs problèmes qui empêchent l’implantation des robots parallèles à câbles en milieu industriel nécessitent encore un peu de travail. Le but principal de ce manuscrit est d’étudier et de proposer des solutions à quelques problèmes qui pourraient survenir lors de la mise en service de tels robots. Premièrement, une méthode d’identification du couple d’encochage aux moteurs est utilisée afin d’uniformiser le mouvement à basse vitesse des moteurs. Les courbes de couple d’encochage résultantes sont par la suite incorporées à un schéma de commande qui est utilisé de manière subséquente. Deuxièmement, des essais de téléopération sont effectués avec un robot plan afin de simuler une tâche d’assemblage à moyenne échelle. Ces tests permettent de déterminer qu’une précision de ±5.3 mm est atteignable lorsqu’un humain commande l’effecteur en étant basé sur une rétroaction visuelle seulement. Troisièmement, on remarque généralement que les robots parallèles entraînés par câbles ont une précision absolue moindre que les autres robots parallèles. Afin de tenter de résoudre ce problème, une méthode d’étalonnage automatique à l’aide d’une caméra embarquée est proposée. Celle-ci inclut une méthode de génération d’une liste de poses d’étalonnage. Un modèle d’erreur est également proposé et comparé aux résultats obtenus en simulation. En dernier lieu, une validation expérimentale est réalisée, puis une vérification permet de conclure que l’application de la méthode se traduit par une erreur moyenne de 31 mm en position et 1.6° en orientation. / Cable-driven parallel robots are a type of manipulator allowing for a large translational workspace and superior dynamic capabilities than their traditional equivalent. Nevertheless, these robots still remain little used and the few prototypes that exist are mostly found within academia. Several problems preventing the implementation of cable-driven parallel robots in industrial environments still need to be worked on. The purpose of this manuscript is to study and propose solutions to some problems that may arise during the putting into service of such robot. First, to ensure uniform operation of the robot at low speed, an identification of the cogging torque is performed. Resulting cogging torque curves are then incorporated into a control scheme that is subsequently used. Second, an accuracy assessment is carried out on a planar robot in order to mimic a mediumsized assembly task. These tests determine that an accuracy of ±5.3mm can be achieved when a human teleoperatates the end effector using only a visual feedback. Third, cable-driven parallel robots are generally less accurate in absolute than other parallel robots. In an effort to solve this issue, an automatic eye-on-hand calibration method is proposed. It includes an algorithm for the generation of a list of calibration poses. An error model is also proposed and the results are compared to the ones obtained by simulation. Finally, an experimental validation is carried out along with a verification of the resulting accuracy. The mean post-calibration error measured is 31 mm in position and 1.6° in orientation.

The computed-torque algorithm is a popular model-based robot trajectory control scheme. Adding an adaptive mechanism to this scheme can improve the error tracking capabilities of the robot controller. This thesis describes the algorithms necessary to develop a computer simulation for the PUMA 560 robot arm. Several robot controllers are outlined with an emphasis on the computed-torque scheme. The PUMA simulation is used to analyze the error tracking capabilities of an adaptive computed-torque controller. Consideration is given to parameter mismatch, unmodeled friction, unknown loading, and path excitation. This thesis shows that even with inaccurate load knowledge the adaptive algorithm enhances the tracking capabilities of the controller.

Ingeniero Civil Eléctrico / El objetivo general del presente trabajo es el de implementar un sistema de cuadricopteros miniatura controlados de forma autónoma, en un ambiente cooperativo, para el transporte de algún tipo de material de forma acumulativa. Se busca estudiar los posibles rangos de aplicaciones que puedan desarrollar cuadricopteros de tamaño reducido, que no sean factibles para drones de mayor tamaño, en cuanto a su desempeño, movilidad y escalabilidad. El sistema consiste en cuadricopteros miniatura Crazyflie controlados utilizando teoría de control PID. Para realizar el feedback del lazo cerrado de control, se utilizara un sistema de captura de movimiento OptiTrack, un conjunto de cámaras para obtener la ubicación espacial de los cuadricopteros y todos los objetos pertinentes en la escena. El sistema de OptiTrack funciona junto a su software Motive:Tracker, el cual puede hacer transmisión de la información de la escena en tiempo real. Se escribe un programa en lenguaje de programación Python para recibir e interpretar toda la información recibida de la escena, implementar el control PID y crear una rutina de comportamiento para los cuadricopteros. Respecto al hardware, ha de configurarse el set de cámaras, ensamblar los cuadricopteros y diseñar el mecanismo de sujeción que utilizaran para para transportar carga. Se logra implementar el sistema, desarrollando una iteración en el control de los cuadricopteros para adecuarse al propósito de transporte de material y trabajar acorde a las condiciones de su ambiente. Se consigue levantar 7gr de peso adicional, considerando que un cuadricoptero pesa 19gr. De la misma forma, el diseño del transporte de carga se ha iterado hasta lograr un funcionamiento adecuado, moviendo aproximadamente 2.7gr de arena por viaje.