Vision, as a key perceptional capability for cognitive systems relates to rather difficult problems -such as visual object recognition, representation, categorization, and scene understanding. State-of-the-art solutio...
详细信息
Vision, as a key perceptional capability for cognitive systems relates to rather difficult problems -such as visual object recognition, representation, categorization, and scene understanding. State-of-the-art solutions, using object appearance based models, already reached certain maturity. they achieve excellent recognition performance and provide learning structures that are subsequently utilized for object recognition and tracking. However, in context of object topology understanding for cognitive tasks, these methods cannot be directly compared with human performance, because it is obvious that appearance based methods do not contribute to understanding of structures in 3D. Research findings from infant psychology and animal investigation give evidence for using hierarchical models of object representation, based on image primitives e.g. such as edges, corners, shading or homogeneity of object colors. It is the objective of this paper to present an approach based on both, findings from biological studies and cognitive science, as enablers for autonomous cognitive investigation of natural scenes and their understanding. We present the architecture of a compound cognitive framework and its first behavioral level withthe implementation of a vision model of the mammalian striate visual cortex in five layers. the proposed implementation is exemplified with an object similar to the Necker cube.
the main objective of a standard reinforcement learner is usually defined as maximization of a scalar reward function given externally from the environment. On the other hand, an intrinsically motivated reinforcement ...
详细信息
the main objective of a standard reinforcement learner is usually defined as maximization of a scalar reward function given externally from the environment. On the other hand, an intrinsically motivated reinforcement learner creates an intrinsic reward function from its own criteria such as curiosity, prediction error, and learning progress. this paper proposes a novel approach to deal with both intrinsic and extrinsic rewards for reinforcement learning from a viewpoint of constrained optimization problem. the extrinsic rewards construct inequality constraints to the stochastic policy while the intrinsic reward determines the current objective function for the learning agent. By integrating policy gradient reinforcement learning algorithms and techniques used in nonlinear programming, our proposed method, named the constrained policy gradient reinforcement learning (CPGRL), maximizes the long-term average intrinsic reward under the inequality constraints induced by the extrinsic rewards. the CPGRL is successfully applied to a simple MDP problem and a control task of a robot arm.
It is well known that the massage therapy is useful for the rehabilitation of various diseases. Although various apparatus have been developed for the massage of the torso and limbs, a machine to perform precise massa...
详细信息
It is well known that the massage therapy is useful for the rehabilitation of various diseases. Although various apparatus have been developed for the massage of the torso and limbs, a machine to perform precise massage therapy to maxillofacial region is not developed yet. therefore, we developed a robot system that provides massage therapy to maxillofacial region. the name of our newly developed robot was WAO-1. WAO-1 has been designed to perform appropriate massage to the patients with dry mouth, etc. WAO-1 is composed by two 6-degree of freedom arms with plungers attached at the end-effectors. the massage is applied to the patient by controlling the force and position of the plunger (virtual compliance). As a preliminary step of the clinical application, a simulation therapy of robotic massage to stimulate salivary flow from parotid gland was performed. the dynamic force on patient's head during the massage using WAO-1 was evaluated. the results suggested that WAO-1 could perform equivalent massage to human hand. We expect that WAO-1 will be useful to provide massage therapy to several patients with oral health problems. this will aim in reducing the therapy costs due to the need of human efforts when long-time therapies are needed.
We present our design and online experiments of a 3-class asynchronous BCI controlling a simulated robot in an indoor environment. Two characteristics of our design have efficiently decreased the false positive rate d...
详细信息
We present our design and online experiments of a 3-class asynchronous BCI controlling a simulated robot in an indoor environment. Two characteristics of our design have efficiently decreased the false positive rate during the NC (No control) mode. First, three one-vs-rest LDA classifiers are combined to controlthe switching between NC and IC (In control) mode. Second, the hierarchical structure of our controller allows the most reliable class (mental task) in a specific subject to play a dominant role in the robot control. A group of simple rules triggered by local sensor signals are designed for safety and obstacle avoidance in the NC mode. In online experiments, subjects successfully controlled the robot to circumnavigate obstacles and reach small targets in separate rooms.
In this paper we introduce Dynamics Based control (DBC), an approach to planning and control of an agent in stochastic environments. Unlike existing approaches, which seek to optimize expected rewards (e.g., in Partia...
详细信息
ISBN:
(纸本)9788190426275
In this paper we introduce Dynamics Based control (DBC), an approach to planning and control of an agent in stochastic environments. Unlike existing approaches, which seek to optimize expected rewards (e.g., in Partially Observable Markov Decision Problems (POMDPs)), DBC optimizes system behavior towards specified system dynamics. We show that a recently developed planning and control approach, Extended Markov Tracking (EMT) is an instantiation of DBC. EMT employs greedy action selection to provide an efficient control algorithm in Markovian environments. We exploit this efficiency in a set of experiments that applied multi-target EMT to a class of area-sweeping problems (searching for moving targets). We show that such problems can be naturally defined and efficiently solved using the DBC framework, and its EMT instantiation.
Modem robotics today deals with increasing requirements on the flexible automation. One of this is the usage of linear tracks or even called 7(th) axis to extend the robots workspace. the inaccuracies of the linear tr...
详细信息
ISBN:
(纸本)9728865600
Modem robotics today deals with increasing requirements on the flexible automation. One of this is the usage of linear tracks or even called 7(th) axis to extend the robots workspace. the inaccuracies of the linear track deteriorate the accuracy, which is in constrast to highly accurate robot systems needed for modem applications. To enhance the accuracy of the system consisting of robot and linear track, an identification of the nonlinearities of the linear track is necessary. this article introduces an optimisation of a method for highly accurate integration of track motions where the profile of the linear track is identified by single coordinate systems along the track, combined by a cubic spline interpolation. Resulting there is a continous description of the track profile, depending on the current position of the robot on the linear track.
Networked robotics is an area that straddles robotics and network technology. A robot system controlled via the WWW exploits the Internet network and hence is one realisation of networked robotics. A set of field robo...
详细信息
ISBN:
(纸本)9728865600
Networked robotics is an area that straddles robotics and network technology. A robot system controlled via the WWW exploits the Internet network and hence is one realisation of networked robotics. A set of field robots that exploit wireless networks to share and distribute tasks might also be considered an exemplar of networked robotics. But isn't this just an exemplar of distributed robotics? And if so, what does networked robotics bring to the 'robotics' table? these are questions and issues addressed in this paper. the paper will propose that networks are at once both enabling and constraining to robotics. they enlarge the scope of the robotics discipline yet introduce challenges that must be overcome if that potential is to be fully realized. In short, when the network becomes a design issue - normally when performance of the system is at a premium - networked robotics is at play.
this paper tests an implementation of the data driven Alarkov chain Alonte Carlo image segmentation procedure used by TO and Zhu. the test we use is a new emprical test based on the ability of an application to make c...
详细信息
ISBN:
(纸本)9781424403417
this paper tests an implementation of the data driven Alarkov chain Alonte Carlo image segmentation procedure used by TO and Zhu. the test we use is a new emprical test based on the ability of an application to make certain types of modifications to the image based on the segmentation results.
the accuracy of a pose measurement system, called PosEye, is investigated. PosEye is a system for real time measurement of the position and orientation, the pose, of a camera (sensor) using the information in its imag...
详细信息
ISBN:
(纸本)9781424403417
the accuracy of a pose measurement system, called PosEye, is investigated. PosEye is a system for real time measurement of the position and orientation, the pose, of a camera (sensor) using the information in its image. this sensor is aimed to be mounted on an industrial robot for welding. the investigation was done by comparing the PosEye system position output to that of a coordinate measuring machine. Sources of errors are identified, and suggestions for improvements are made.
this paper presents a new strategy for robust tracking in robot manipulators. the aim of the strategy is to reject parametric uncertainties due to model or load disturbances. the basic controller acting on the manipul...
详细信息
ISBN:
(纸本)9728865600
this paper presents a new strategy for robust tracking in robot manipulators. the aim of the strategy is to reject parametric uncertainties due to model or load disturbances. the basic controller acting on the manipulator is a robust controller designed by Lyapunov's direct method. Acting on this controller there is an adaptive system responsible for the adaptation of the basic parameter of the robust feedforward term. the performance of the strategy is tested in a Puma-560 manipulator. A comparison with existing techniques is done to verify the efficiency of the presented controller.
暂无评论