Explore chapters and articles related to this topic
Intelligent Robotic Vision Systems
Published in Spyros G. Tzafestas, Intelligent Robotic Systems, 2020
L. Van Gool, P. Wambacq, A. Oosterlinck
In visual servoing applications, the vision system’s output is not used to select a program with predefined actions or to set some parameters but instead is used to adaptively control the robot’s motions in feedback mode. These integrated systems have the potential for improving robot accuracy and immunity against environmental changes. In sensorless or open-loop sensor-controlled robot configurations, the control loop is not closed around the end-effector position. This leads to inherent limitations in compensating for inaccurate arm models, cumulative errors in the path of autonomous robots, and unpredictable changes in object positions and orientations. When the sensor is placed on the robot itself, real-time updates of strategies according to changing environments or renewed localization of reference points is possible. Two basically different feedback modes are the static and the dynamic.
Introduction
Published in Laxmidhar Behera, Swagat Kumar, Prem Kumar Patchaikani, Ranjith Ravindranathan Nair, Samrat Dutta, Intelligent Control of Robotic Systems, 2020
Laxmidhar Behera, Swagat Kumar, Prem Kumar Patchaikani, Ranjith Ravindranathan Nair, Samrat Dutta
Vision-based control of a redundant manipulator is a challenging task which involves two sub-tasks: (i) visual servoing [10, 11] and (ii) redundant manipulator control [12, 13]. In general, visual servoing computes the end-effector velocity required to reach the desired position from the image features obtained through the visual feedback. It basically assumes that there exists a non-redundant manipulator which can generate the desired end-effector velocity with its own inverse kinematic algorithm. A non-redundant manipulator can achieve the end-effector velocity estimated from visual servoing with unique joint angle configuration. In a dynamic environment, the available unique joint angle configuration may become infeasible to position the end-effector due to the presence of obstacles and the physical constraints. This necessitates the use of redundant manipulators for vision-based control in dynamic environments, which have excess DOF than that required for the given task. Theoretically infinite choices of joint angle configuration exist for redundant manipulators to achieve the estimated end-effector velocity. The excess DOF can be effectively utilized in performing additional constraints introduced by the dynamic environment. An optimal joint angle configuration needs to be selected, while satisfying these additional constraints. This is popularly known as redundancy resolution.
Robotics for Spatially and Temporally Unstructured Agricultural Environments
Published in Dan Zhang, Bin Wei, Robotics and Mechatronics for Agriculture, 2017
Konrad Ahlin, Brad Bazemore, Byron Boots, John Burnham, Dellaert Frank, Jing Dong, Ai-Ping Hu, Benjamin Joffe, Gary McMurray, Glen Rains, Nader Sadegh
Visual servoing is the process of controlling a robotic device using real-time visual information in a feedback loop. Image Based Visual Servoing (IBVS) is a classical approach to visual servoing in robotics that attempts to converge on an object in 3D space by only using information about the objects 2D pixel geometry. IBVS does not assume any information about the object being viewed; instead, this method uses a given set of desired points in the image space that it uses for its error calculations (Chaumette and Hutchinson, 2006). IBVS is a widely studied topic that is often used in robotics applications.
Hardware-in-the-loop testing of current cycle feedback ILC for stabilisation and tracking control of under-actuated visual servo system
Published in International Journal of Systems Science, 2021
Vimala Kumari Jonnalagadda, Vinodh Kumar Elumalai
Visual servoing, which utilises the feedback information obtained from cameras to control the robot/mechanical systems, is the fusion of results from several engineering domains ranging from real time computing, image processing, control theory, kinematics and dynamics. The factors which attract visual feedback in control implementation are non-invasive measurement, versatility and accuracy. Some of the real world applications of visual servoing are autonomous vehicles (Allibert et al., 2019), factory automation (Anis et al., 2008; Chang, 2018), robot manipulator (Dong & Zhang, 2020; S. Liu & Dong, 2020), road-traffic control (Kumaran et al., 2019) and surveillance (Abdessameud & Janabi-Sharifi, 2015; Shirzadeh et al., 2017). The ball on plate system, which is the two-dimensional extension of the ball on beam system, is a typical benchmark visual servoing experimental setup used in several universities to teach and explore the various challenges of vision based control system (Fan et al., 2004). The ball on plate system is an open loop unstable, under-actuated system with nonlinear dynamics and strong inter-axis coupling (Armendariz et al., 2010; Park & Lee, 2003). Hence, designing a control algorithm for precision trajectory tracking is always a challenging task. From control perspective, some of the notable results reported in the literature to solve the stabilisation and tracking control problem of ball on plate system are as follows.
Generalization of reference filtering control strategy for 2D/3D visual feedback control of industrial robot manipulators
Published in International Journal of Computer Integrated Manufacturing, 2022
J. Ernesto Solanes, Pau Muñoz-Benavent, Leopoldo Armesto, Luis Gracia, Josep Tornero
In general, any industrial visual servoing system has, at least, the following main elements (see Figure 1): a vision system in charge of extracting the required information from the environment, coined as visual features vector; an external controller, usually a PLC or PC-based industrial workstation, where the visual control algorithms and other auxiliary algorithms (e.g. communications with the factory servers) are implemented; an industrial robot; and other devices, such as screens or factory displays. All of these elements are connected through an industrial router.
Visually Guided Manipulator Based on Artificial Neural Networks
Published in IETE Journal of Research, 2018
Yasaman Ghandi, Mohsen Davoudi
Visual servoing is a robot control method in which visual feedback or acquired image from the camera is introduced into the control loop directly and could enhance manipulator performance [1]. A manipulator with vision sensor could track and grasp unknown objects or assemble objects and avoid obstacles. Based on image features, visual servoing is classified into two categories: image-based visual servoing (IBVS) and position-based visual servoing (PBVS).