Robotics is required to provide effective solutions for an increasing assortment of applications. Since electronics is cheaper and cheaper, there is a general trend towards systems that are on the one hand computationally demanding, on the other mechanically as simple and inexpensive as possible. In order to meet economic and ecologic specifications, robotic systems must also possess a high degree of modularity and versatility, and be energetically efficient. Cable-driven parallel robots (CDPRs), which use multiple cables to control the end-effector (EE) pose, provide promising solutions to meet the above requirements. They provide peculiar advantages, such as light weight, large workspaces, reduced manufacturing and maintenance costs, ease of assembly, superior modularity and reconfigurability. With the same limited set of cables and actuators, a vast possibility of manipulation tasks may be obtained by simply reconfiguring the cable anchor points and the adopted number of cables/actuators, confining most task-dependent customizations to the EE design. The aim of the project is to develop, and validate on a demonstrator, new methodologies and algorithms for the design and control of a novel family of robots, which we call Intelligent CDPRs (ICABOTs). ICABOTs are supposed to quickly adapt their geometry to the task at hand and to the operation environment, before and during task execution. The development of the project has to go through the study of several fundamental problems, including mechanical design, dynamic modeling and control, and will deliver a modular mechanical architecture, a design software and a control algorithm. The ICABOT architecture will include cable-coiling devices attached on beams of modular frames or installed on movable shuttles for greater versatility, with cables being actuated or not depending on the particular pose and dynamics of the EE. The possibility of connecting cables to each other on variable anchor points, thus creating a distributed web-architecture, will also be explored. The design tool will allow one to determine the best robot geometry, in a broad sense, for the application being considered. The control algorithm will adapt in real-time the input commands and the robot geometry delivered by the design tool, modifying cable configuration and tension distribution to gain the highest performance and/or the lowest energy consumption during the execution of the motion task. The mechanical architecture, the design tool and the control software will be validated by implementing and testing a demonstrator, which will be delivered by the project. The ICABOT technology may spawn a new generation of adaptive robotic systems. Potential applications are in several fields of robotics, particularly those in which low cost, energy saving and versatility have a major impact, including rehabilitation, health, service to humans, domotics, entertainment, rescue operations, and collaborative industrial tasks.
Intelligent Cable-Driven Robots (ICABOT): an adaptive approach to robot design and control (un approccio adattativo alla progettazione ed al controllo dei robot Intelligent)
marco carricato
Supervision
2017
Abstract
Robotics is required to provide effective solutions for an increasing assortment of applications. Since electronics is cheaper and cheaper, there is a general trend towards systems that are on the one hand computationally demanding, on the other mechanically as simple and inexpensive as possible. In order to meet economic and ecologic specifications, robotic systems must also possess a high degree of modularity and versatility, and be energetically efficient. Cable-driven parallel robots (CDPRs), which use multiple cables to control the end-effector (EE) pose, provide promising solutions to meet the above requirements. They provide peculiar advantages, such as light weight, large workspaces, reduced manufacturing and maintenance costs, ease of assembly, superior modularity and reconfigurability. With the same limited set of cables and actuators, a vast possibility of manipulation tasks may be obtained by simply reconfiguring the cable anchor points and the adopted number of cables/actuators, confining most task-dependent customizations to the EE design. The aim of the project is to develop, and validate on a demonstrator, new methodologies and algorithms for the design and control of a novel family of robots, which we call Intelligent CDPRs (ICABOTs). ICABOTs are supposed to quickly adapt their geometry to the task at hand and to the operation environment, before and during task execution. The development of the project has to go through the study of several fundamental problems, including mechanical design, dynamic modeling and control, and will deliver a modular mechanical architecture, a design software and a control algorithm. The ICABOT architecture will include cable-coiling devices attached on beams of modular frames or installed on movable shuttles for greater versatility, with cables being actuated or not depending on the particular pose and dynamics of the EE. The possibility of connecting cables to each other on variable anchor points, thus creating a distributed web-architecture, will also be explored. The design tool will allow one to determine the best robot geometry, in a broad sense, for the application being considered. The control algorithm will adapt in real-time the input commands and the robot geometry delivered by the design tool, modifying cable configuration and tension distribution to gain the highest performance and/or the lowest energy consumption during the execution of the motion task. The mechanical architecture, the design tool and the control software will be validated by implementing and testing a demonstrator, which will be delivered by the project. The ICABOT technology may spawn a new generation of adaptive robotic systems. Potential applications are in several fields of robotics, particularly those in which low cost, energy saving and versatility have a major impact, including rehabilitation, health, service to humans, domotics, entertainment, rescue operations, and collaborative industrial tasks.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
