一种基于联合学习的家庭日常工具功用性部件检测算法

对工具及其功用性部件的认知是共融机器人智能提升的重要研究方向.本文针对家庭日常工具的功用性部件建模与检测问题展开研究,提出了一种基于条件随机场（Conditional random field,CRF）和稀疏编码联合学习的家庭日常工具功用性部件检测算法.首先,从工具深度图像提取表征工具功用性部件的几何特征;然后,分析CRF和稀疏编码之间的耦合关系并进行公式化表示,将特征稀疏化后作为潜变量构建初始条件随机场模型,并进行稀疏字典和CRF的协同优化：一方面,将特征的稀疏表示作为CRF的随机变量条件及权重参数选择器;另一方面,在CRF调控下对稀疏字典进行更新.随后使用自适应时刻估计（Adaptive moment estimation,Adam）方法实现模型解耦与求解.最后,给出了基于联合学习的工具功用性部件模型离线构建算法,以及基于该模型的在线检测方法.实验结果表明,相较于使用传统特征提取和模型构建方法,本文方法对功用性部件的检测精度和效率均得到提升,且能够满足普通配置机器人对工具功用性认知的需要.     

A domain-specific language for context modeling in context-aware systems

Context-awareness refers to systems that can both sense and react based on their environment. One of the main difficulties that developers of context-aware systems must tackle is how to manage the needed context information. In this paper we present MLContext, a textual Domain-Specific Language (DSL) which is specially tailored for modeling context information. It has been implemented by applying Model-Driven Development (MDD) techniques to automatically generate software artifacts from context models. The MLContext abstract syntax has been defined as a metamodel, and model-to text transformations have been written to generate the desired software artifacts. The concrete syntax has been defined with the EMFText tool, which generates an editor and model injector.     

Implicit human computer interaction through context

In this paper the term implicit human-computer interaction is defined. It is discussed how the availability of processing power and advanced sensing technology can enable a shift in HCI from explicit interaction, such as direct manipulation GUIs, towards a more implicit interaction based on situational context. In the paper, an algorithm is given based on a number of questions to identify applications that can facilitate implicit interaction. An XML-based language to describe implicit HCI is proposed. The language uses contextual variables that can be grouped using different types of semantics as well as actions that are called by triggers. The term of perception is discussed and four basic approaches are identified that are useful when building context-aware applications. Two examples, a wearable context awareness component and a sensor-board, show how sensor-based perception can be implemented. It is also discussed how situational context can be exploited to improve input and output of mobile devices.     

Bayesian approach to sensor-based context awareness

The usability of a mobile device and services can be enhanced by context awareness. The aim of this experiment was to expand the set of generally recognizable constituents of context concerning personal mobile device usage. Naive Bayesian networks were applied to classify the contexts of a mobile device user in her normal daily activities. The distinguishing feature of this experiment in comparison to earlier context recognition research is the use of a naive Bayes framework, and an extensive set of audio features derived partly from the algorithms of the upcoming MPEG-7 standard. The classification was based mainly on audio features measured in a home scenario. The classification results indicate that with a resolution of one second in segments of 5–30 seconds, situations can be extracted fairly well, but most of the contexts are likely to be valid only in a restricted scenario. Naive Bayes framework is feasible for context recognition. In real world conditions, the recognition accuracy using leave-one-out cross validation was 87% of true positives and 95% of true negatives, averaged over nine eight-minute scenarios containing 17 segments of different lengths and nine different contexts. Respectively, the reference accuracies measured by testing with training data were 88% and 95%, suggesting that the model was capable of covering the variability introduced in the data on purpose. Reference recognition accuracy in controlled conditions was 96% and 100%, respectively. However, from the applicability viewpoint, generalization remains a problem, as from a wider perspective almost any feature may refer to many possible real world situations.     

Architecture for management and fusion of context information

Information in a context-aware system has diverse natures. Raw data coming from sensors are aggregated and filtered to create more abstract information, which can be processed by context-aware application components to decide what actions should be performed. This process involves several activities: finding the available sources of information and their types, gathering the data from these sources, facilitating the fusion (aggregation and interpretation) of the different pieces of data, and updating the representation of the context to be used by applications. The reverse path also appears in context-aware systems, from changes in the context representation to trigger actions in certain actuators. FAERIE (Framework for AmI: Extensible Resources for Intelligent Environments) is a framework that facilitates management and fusion of context information at different levels. It is implemented as a distributed blackboard model. Each node of the system has a private blackboard to manage pieces of information that can be accessed by observer components, either locally or remotely (from other nodes) in a transparent way. The use of the framework is illustrated with a case study of an application for guiding people to meetings in a university building.     

Acoustic environment as an indicator of social and physical context

Acoustic environments provide many valuable cues for context-aware computing applications. From the acoustic environment we can infer the types of activity, communication modes and other actors involved in the activity. Environmental or background noise can be classified with a high degree of accuracy using recordings from microphones commonly found in PDAs and other consumer devices. We describe an acoustic environment recognition system incorporating an adaptive learning mechanism and its use in a noise tracker. We show how this information is exploited in a mobile context framework. To illustrate our approach we describe a context-aware multimodal weather forecasting service, which accepts spoken or written queries and presents forecast information in several forms, including email, voice and sign languages.

A resource and context model for mobile middleware

Mobile computing systems should be self-managed to simplify operation and maintenance plus meet user’s expectation with respect to Quality of Service (QoS). When architecting self-managed mobile computing systems, one must take a holistic view on both QoS management and the entities in the mobile environment. This paper presents a novel model that includes both resources and context elements. To illustrate the usefulness of the model, it is applied to a video streaming application by: (1) modelling context elements and resources in the environment, (2) specifying context dependencies and QoS characteristics of the application, and (3) designing weakly integrated resource and context managers. We describe a middleware that uses the developed managers when evaluating context dependencies and predict offered QoS of alternative implementations of the application. In order to select the one that can operate in the current environment and that best satisfies given user preferences.     

A context awareness framework for cross-platform distributed applications

With the introduction of interconnected cross-platform middleware, a new area of opportunities for ubiquitous/pervasive computing has emerged. Context aware applications can be enhanced to practically and realistically incorporate multiple facets of human–machine interactions in everyday life that are not limited to a device-centered model for deducing context. In this paper, we propose that they can rather extend this model to a human-centered, device and platform independent model, based on a personal distributed application and data cloud ecosystem. For this to be achieved, webinos, a set of web runtime extensions that enable web applications and services to be used and shared consistently and securely over a broad spectrum of converged and connected devices, is used to provide this ecosystem. The webinos Context Awareness Framework described here is accessible to each webinos-enabled application. After strict policy enforcement, it can collect contextual information, either via an automatic mechanism that intercepts native calls made by webinos applications through the various webinos APIs, via an automatic polling mechanism to these APIs, or via custom, application-specific context schema extensions. It can then distribute the contextual information from its own personal cloud storage mechanism, in the form of simple, manageable and intuitive Context Objects, to and from all webinos-enabled devices owned by the same user, or even other, authorized users.     

Dealing with the effects of context mismatch in group work

The context concept can be used with advantage in the area of Computer-Supported Cooperative Work. In many studies, several forms of context have been used without explicit association to the context concept. This paper attempts to clarify the relationship between context and group work. In particular, a framework is proposed to understand context as connected to other concepts used in group work. The framework is useful to analyze some groupware systems from the context perspective and to obtain some insight on possible improvements for users.     

Transports of delight? What the experience of receiving (mobile) phone calls can tell us about design

This paper takes a phenomenological approach to analysing people’s accounts of receiving phone calls, drawing on Heidegger and Feenberg. Accounts of calls received on a mobile phone are compared with those on landlines, charting progress from location-centred to person-centred phoning. A range of naturally-occurring contexts are discussed in terms of the experience of balancing the activities of talking on the phone with activities in the immediate environment, and the enchantment sustained or sacrificed. The study suggests that recipients’ enchantment with phoning is affected by their freedom and desire to project towards the caller and create shared spaces, and reveals some factors that impact on the transitions of attention required to do so. It concludes with the design implications of taking this view of interactions with and through phones.     

Network aware dynamic context subscription management

Context-awareness is a key requirement in many of today’s networks, services and applications. Context management systems are used to provide access to distributed, dynamic context information. The reliability of remotely accessed dynamic context information is impacted by network delay, packet drop probability, its information dynamics and the access strategy used. Due to the characteristics of the different access strategies, different levels of reliability of context information can be ensured, but at the same time, these strategies lead to different access traffic which impacts also the network performance, and hence feeds back to the reliability of the information. Furthermore, different levels of QoS may be available and used in order to mitigate the impact of network performance degradation on the reliability of the dynamic context information. In this paper we describe a system and algorithms that are capable of configuring effectively context access strategies in order to maximize reliability of all accessed dynamic context information. The framework utilizes and extends existing information reliability models, and it can utilize different network performance models. Simulation results of scenarios in which the framework uses finite-buffer bottleneck performance models demonstrate the effectiveness of our algorithm to increase reliability. Furthermore, the framework is applied to a scenario with QoS classes that allows to trade off delay and loss via different buffer-size configurations.     

Context unification with one context variable

The context unification problem is a generalization of standard term unification. It consists of finding a unifier for a set of term equations containing first-order variables and context variables. In this paper we analyze the special case of context unification where the use of at most one context variable is allowed and show that it is in NP. The motivation for investigating this subcase of context unification is interprocedural program analysis for programs described using arbitrary terms, generalizing the case where terms were restricted to using unary function symbols. Our results imply that the redundancy problem is in coNP, and that the finite redundancy property holds in this case. We also exhibit particular cases where one context unification is polynomial.     

ML aided context feature extraction for cognitive radio

This paper addresses the estimation of different context features of a primary user network, such as transmitters’ positions, antenna patterns and directions, and propagation model characteristics. It is based on radio signal strength measurements obtained by a sensor network without any prior knowledge about the configuration of the primary transmitters in terms of antenna types or propagation model. A Maximum Likelihood Aided Context Feature Extraction (MLACFE) method is introduced based on applying image processing and a Maximum Likelihood estimation algorithm over the set of measurements to identify the existing transmitters in the scenario and their parameters. The proposed method can provide a quite similar performance than a classical ML method, in terms of average estimation errors while at the same time reducing the computation time in about three orders of magnitude, for the considered case study.     

Non-parametric hand pose estimation with object context

In the spirit of recent work on contextual recognition and estimation, we present a method for estimating the pose of human hands, employing information about the shape of the object in the hand. Despite the fact that most applications of human hand tracking involve grasping and manipulation of objects, the majority of methods in the literature assume a free hand, isolated from the surrounding environment. Occlusion of the hand from grasped objects does in fact often pose a severe challenge to the estimation of hand pose. In the presented method, object occlusion is not only compensated for, it contributes to the pose estimation in a contextual fashion; this without an explicit model of object shape. Our hand tracking method is non-parametric, performing a nearest neighbor search in a large database (.. entries) of hand poses with and without grasped objects. The system that operates in real time, is robust to self occlusions, object occlusions and segmentation errors, and provides full hand pose reconstruction from monocular video. Temporal consistency in hand pose is taken into account, without explicitly tracking the hand in the high-dim pose space. Experiments show the non-parametric method to outperform other state of the art regression methods, while operating at a significantly lower computational cost than comparable model-based hand tracking methods.     

Important context changes for talking and text messaging during homeward commutes

This paper models the suitability of homeward commutes as a context for talking on a mobile telephone and text messaging. Analysis of these models identifies when and where large changes in suitability frequently arise. To bring commuters the greatest benefits, these are the changes upon which future applications of context-sensitivity and awareness need to focus.     

Important context changes for talking and text messaging during homeward commutes

This paper models the suitability of homeward commutes as a context for talking on a mobile telephone and text messaging. Analysis of these models identifies when and where large changes in suitability frequently arise. To bring commuters the greatest benefits, these are the changes upon which future applications of context-sensitivity and awareness need to focus.     

Visualizing and exploring profiles with calling context ring charts

Calling context profiling is an important technique for analyzing the performance of object‐oriented software with complex inter‐procedural control flow. The Calling Context Tree (CCT) is a common data structure that stores dynamic metrics, such as CPU time, separately for each calling context. As CCTs may comprise millions of nodes, there is a need for a condensed visualization that eases the localization of performance bottlenecks. In this article, we discuss Calling Context Ring Charts (CCRCs), a compact visualization for CCTs, where callee methods are represented in ring segments surrounding the caller's ring segment. In order to reveal hot methods, their callers, and callees, the ring segments can be sized according to a chosen dynamic metric. We describe two case studies where CCRCs help us to detect and fix performance problems in applications. A performance evaluation also confirms that our implementation can efficiently handle large CCTs. Copyright © 2010 John Wiley & Sons, Ltd.     

Learning deterministic context free grammars: The Omphalos competition

This paper describes the winning entry to the Omphalos context free grammar learning competition. We describe a context-free grammatical inference algorithm operating on positive data only, which integrates an information theoretic constituent likelihood measure together with more traditional heuristics based on substitutability and frequency. The competition is discussed from the perspective of a competitor. We discuss a class of deterministic grammars, the Non-terminally Separated (NTS) grammars, that have a property relied on by our algorithm, and consider the possibilities of extending the algorithm to larger classes of languages. Editor: Georgios Paliouras and Yasubumi Sakakibara     

Afforded actions as a behavioral assessment of physical presence in virtual environments

A particular affordance was used as a potential candidate for behavioral assessment of physical presence in virtual environments. The subjects’ task was to walk through a virtual aperture of variable widths. In the case of presence, the subjects’ body orientation, while walking, was hypothesized to be adapted to the width of the aperture and to their own shoulder width. Results show that most subjects adapted their behavior to both their body architecture and the virtual width constraints. These subjects exhibited a behavioral transition from frontal walking to body rotation while walking through broad to narrow apertures. The same behavioral transition has already been documented in real environments (Warren and Whang in J Exp Psychol Human Percept Perform 13(3):371–383, 1987). This behavioral adjustment is thus assumed to be an objective indication of presence. Beyond these results, the present study suggests that every afforded action could be a potential tool for sensorimotor assessment of physical presence. Parts of this research have been first presented at the 11th Annual International Workshop of Presence, Padova, Italy, 16-18 October 2008.