We describe the architecture, algorithms, and experimental testbed for the deployment of large numbers of cooperating robots, and applications to tasks like manipulation and transportation. The coordination between ro...
详细信息
ISBN:
(纸本)9783642147432
We describe the architecture, algorithms, and experimental testbed for the deployment of large numbers of cooperating robots, and applications to tasks like manipulation and transportation. The coordination between robots is completely decentralized to enable scaling up to large numbers of robots. There is no labeling or identification of robots and all robots (and their software) are identical allowing robustness to failures, ease of programming, and modularity enabling the addition or deletion of robots from the team. Our approach requires minimal communication and sensing and the proposed controllers are based only on local information. Moreover, our architecture facilitates asymmetric communication from one or more supervisory agents that can broadcast information to all robots and close the loop by acquiring abstract, high level information related to the supervised robots. We discuss the hardware and software implementation, the architecture, and present recent experimental results.
Modeling and predicting human and vehicle motion is an active research domain. Due to the difficulty of modeling the various factors that determine motion (eg internal state, perception, etc.) this is often tackled by...
详细信息
ISBN:
(纸本)9783642147432
Modeling and predicting human and vehicle motion is an active research domain. Due to the difficulty of modeling the various factors that determine motion (eg internal state, perception, etc.) this is often tackled by applying machine learning techniques to build a statistical model, using as input a collection of trajectories gathered through a sensor (eg camera, laser scanner), and then using that model to predict further motion. Unfortunately, most current techniques use off-line learning algorithms, meaning that they are not able to learn new motion patterns once the learning stage has finished. In this paper, we present an approach which is able to learn new motion patterns incrementally, and in parallel with prediction. Our work is based on a novel extension to Hidden Markov Models called Crowing Hidden Markov models.
The aim of this research is to develop a direct teaching system for multifingered robot hand to reproduce in-hand manipulation demonstrated by an human operator. A recognition method by observing contact state transit...
详细信息
ISBN:
(纸本)9783642147432
The aim of this research is to develop a direct teaching system for multifingered robot hand to reproduce in-hand manipulation demonstrated by an human operator. A recognition method by observing contact state transition on a palm surface is described to detect primitives of in-hand manipulation. Dynamic programming (DP) matching is applied to recognize the primitives. The direct teaching system is developed consisting of an object with multiple sensors and a multi-fingered robot hand "NAIST-hand" developed by our group. By taking a barcode scanning task as an example, an experiment is conducted to demonstrate the validity of the developed system.
This paper presents a stochastic motion planning algorithm and its application to traffic navigation. The algorithm copes with the uncertainty of road traffic conditions by stochastic modeling of travel delay on road ...
详细信息
ISBN:
(纸本)9783642003110
This paper presents a stochastic motion planning algorithm and its application to traffic navigation. The algorithm copes with the uncertainty of road traffic conditions by stochastic modeling of travel delay on road networks. The algorithm determines paths between two points that optimize a cost function of the delay probability distribution. It can be used to find paths that maximize the probability of reaching a destination within a particular travel deadline. For such problems, standard shortest-path algorithms don't work because the optimal substructure property doesn't hold. We evaluate our algorithm using both simulations and real-world drives, using delay data gathered from a set of taxis equipped with GPS sensors and a wireless network. Our algorithm can be integrated into on-board navigation systems as well as route-finding Web sites, providing drivers with good paths that meet their desired goals.
We aim to develop the Hybrid Assistive Lims (HAL) in order to enhance and upgrade the human capabilities based on the frontier science Cybernics. Cybernics is a new domain of interdisciplinary research centered on cyb...
详细信息
ISBN:
(纸本)9783642147432
We aim to develop the Hybrid Assistive Lims (HAL) in order to enhance and upgrade the human capabilities based on the frontier science Cybernics. Cybernics is a new domain of interdisciplinary research centered on cybernetics, mechatronics, and informatics, and integrates neuroscience, robotics, systems engineering, information technology, "kansei" engineering, ergonomics, physiology, social science, law, ethics, management, economics etc. Robot Suit HAL is a cyborg type robot that can expand, augment and support physical capability. The robot suit HAL has two types of control systems such as "Cybernic Voluntary Control System" and "Cybernic Autonomous Control System". The application fields of HAL are medical welfare, heavy work support and entertainment etc. In this paper, the outline of HAL and some of the important algorithms and recent challenges are described.
This work studies the interaction of the nonholonomic and visibility constraints of a robot that has to maintain visibility of a static landmark. The robot is a differential drive system and has a sensor with limited ...
详细信息
ISBN:
(纸本)9783642003110
This work studies the interaction of the nonholonomic and visibility constraints of a robot that has to maintain visibility of a static landmark. The robot is a differential drive system and has a sensor with limited field of view. We determine the necessary and sufficient conditions for the existence of a path for our system to be able to maintain landmark visibility in the presence of obstacles. We present a complete motion planner that solves this problem based on a recursive subdivision of a path computed for a holonomic robot with the same visibility constraints.
This paper presents a general probabilistic framework for multi-sensor multi-class object recognition based on Conditional Random Fields (CRFs) trained with virtual evidence boosting. The learnt representation models ...
详细信息
ISBN:
(纸本)9783642147432
This paper presents a general probabilistic framework for multi-sensor multi-class object recognition based on Conditional Random Fields (CRFs) trained with virtual evidence boosting. The learnt representation models spatial and temporal relationships and is able to integrate arbitrary sensor information by automatically extracting features from data. We demonstrate the benefits of modelling spatial and temporal relationships for the problem of detecting seven classes of objects using laser and vision data in outdoor environments. Additionally, we show how this framework can be used with partially labeled data, thereby significantly reducing the burden of manual data annotation.
In a crystallographic experiment, a protein is precipitated to obtain a crystalline sample (crystal) containing many copies of the molecule. An electron density map (EDM) is calculated from diffraction images obtained...
详细信息
ISBN:
(纸本)9783642003110
In a crystallographic experiment, a protein is precipitated to obtain a crystalline sample (crystal) containing many copies of the molecule. An electron density map (EDM) is calculated from diffraction images obtained from focusing X-rays through the sample at different angles. This involves iterative phase determination and density calculation. The protein conformation is modeled by placing the atoms in 3-D space to best match the electron density. In practice, the copies of a protein in a crystal are not exactly in the same conformation. Consequently the obtained EDM, which corresponds to the cumulative distribution of atomic positions over all conformations, is blurred. Existing modeling methods compute an "average" protein conformation by maximizing its fit with the EDM and explain structural heterogeneity in the crystal with a harmonic distribution of the position of each atom. However, proteins undergo coordinated conformational variations leading to substantial correlated changes in atomic positions. These variations are biologically important. This paper presents a sample-select approach to model structural heterogeneity by computing an ensemble of conformations (along with occupancies) that, collectively, provide a near-optimal explanation of the EDM. The focus is on deformable protein fragments, mainly loops and side-chains. Tests were successfully conducted on simulated and experimental EDMs.
This book is a revised version of the doctoral dissertation presented by D. Ribas in the Department of Computer Engineering at the University of Girona. The main purpose of this work is to present different techniques...
ISBN:
(数字)9783642140402
ISBN:
(纸本)9783642140396
This book is a revised version of the doctoral dissertation presented by D. Ribas in the Department of Computer Engineering at the University of Girona. The main purpose of this work is to present different techniques developed with the objective of providing a solution to the navigation problem for Autonomous Underwater Vehicles (AUVs) operating in structured environments, with special attention to localization techniques but, particularly, to the application of SLAM (Simultaneous Localization And Mapping) techniques as a self-contained system which requires neither previous knowledge of the scenario nor the use of absolute positioning systems like GPS, LBL or USBL. This book also presents techniques for feature extraction capable of dealing with the particular complexities of a mechanically scanned imaging sonar. The approaches described in this book are designed for use in structured environments like those present in many industrial scenarios, specifically for scenarios containing manmade structures in the form of rectilinear walls like those met in harbours, breakwaters, marinas, canal systems, etc. Although most of the previous work done in this field focuses on open sea and coastal applications, obtaining an accurate positioning in such scenarios would increase AUVs capabilities notably, allowing to perform autonomously tasks such as inspection of underwater structures, surveillance of marine installations and even enabling autonomous harbour leaving and returning operations.
Connectivity is an important requirement for wireless sensor networks especially in real-time monitoring and data transfer applications. However, node movements and failures change the topology of the initial deployed...
详细信息
ISBN:
(纸本)9783642003110
Connectivity is an important requirement for wireless sensor networks especially in real-time monitoring and data transfer applications. However, node movements and failures change the topology of the initial deployed network, which can result in partitioning of the communication graph. In this paper, we present a method for maintaining and repairing the communication network of a dynamic mobile wireless sensor network. We assume that we cannot control the motion of wireless sensor nodes, but there are robots whose motion can be controlled by the wireless sensor nodes to maintain and repair the connectivity of the network. At the heart of our method lies a novel graph property, k-redundancy, which is a measure of the importance of a node to the connectivity of a network. We first show that this property can be used to estimate repair time of a dynamic network. Then, we present a dynamic repair algorithm that minimizes expected repair time. Finally, we show the effectiveness of our method with extensive simulations and its feasibility with experiments on real robots and motes.
暂无评论