Papers by Clemens Lombriser
Informatik aktuell, 2007
Context recognition, such as gesture or activity recognition, is a key mechanism that enables ubi... more Context recognition, such as gesture or activity recognition, is a key mechanism that enables ubiquitous computing systems to proactively support users. It becomes challenging in unconstrained environments such as those encountered in daily living, where it has to deal with heterogeneous networks, changing sensor availability, communication capabilities, and available processing power. This paper describes Titan, a new framework that is specifically designed to perform context recognition in such dynamic sensor networks. Context recognition algorithms are represented by interconnected data processing tasks forming a task network. Titan adapts to different context recognition algorithms by dynamically reconfiguring individual sensor nodes to update the network wide algorithm execution. We demonstrate the applicability of Titan for activity recognition on Tmote Sky sensor nodes and show that Titan is able to perform processing of sensor data sampled at 100 Hz and can reconfigure a sensor node in less than 1ms. This results in a better tradeoff between computational speed and dynamic reconfiguration time.
Proceedings of the Second International Conference on Body Area Networks BodyNets, 2007
Recognizing user activities using body-worn, miniaturized sensor nodes enables wearable computers... more Recognizing user activities using body-worn, miniaturized sensor nodes enables wearable computers to act contextaware. This paper describes how online activity recognition algorithms can be run on the SensorButton, our miniaturized wireless sensor platform. We present how the activity recognition algorithms have been optimized to be run online on our sensor platform, and how the execution can be distributed to the wireless sensor network. The resulting algorithm has been implemented as a custom, platform-specific executable as well as integrated into TinyOS. A comparison shows that the TinyOS executable is using about 7kB more code memory, while both implementations classify the activity in up to 18 classifications per second.
Lecture Notes in Computer Science, 2009
Abstract. Quality of Context (QoC) in context-aware computing im-proves reasoning and decision ma... more Abstract. Quality of Context (QoC) in context-aware computing im-proves reasoning and decision making. Activity recognition in wearable computing enables context-aware assistance. Wearable systems must in-clude QoC to participate in context processing frameworks common in ...
2007 3rd International Conference on Intelligent Sensors, Sensor Networks and Information, 2007
This paper presents the e-SENSE middleware architecture for distributed processing of context inf... more This paper presents the e-SENSE middleware architecture for distributed processing of context information in dynamic wireless sensor networks. At the lower layer, the sensor nodes organize into clusters spontaneously, based on a shared context. These clusters form the basis for the service-oriented processing layer, where the functionality of the sensor network is expressed as service task graphs that support the distributed execution of applications. The higher layer is responsible for complex context inference and recognition. As a concrete example we evaluate the distributed recognition of human activities in a car assembly process.
IEEE Journal on Selected Areas in Communications, 2000
Context processing in Body Area Networks (BANs) faces unique challenges due to the user and node ... more Context processing in Body Area Networks (BANs) faces unique challenges due to the user and node mobility, the need of real-time adaptation to the dynamic topological and contextual changes, and heterogeneous processing capabilities and energy constraints present on the available devices. This paper proposes a service-oriented framework for the execution of context recognition algorithms. We describe and theoretically analyze the performance of the main framework components, including the sensor network organization, service discovery, service graph construction, service distribution and mapping. The theoretical results are followed by the simulation of the proposed framework as a whole, showing the overall cost of using dynamically distributed applications on the network.
EURASIP Journal on Wireless Communications and Networking, 2011
Upcoming ambient intelligence environments will boast ever larger number of sensor nodes readily ... more Upcoming ambient intelligence environments will boast ever larger number of sensor nodes readily available on body, in objects, and in the user's surroundings. We envision "Pervasive Apps", user-centric activity-aware pervasive computing applications. They use available sensors for activity recognition. They are downloadable from application repositories, much like current Apps for mobile phones. A key challenge is to provide Pervasive Apps in open-ended environments where resource availability cannot be predicted. We therefore introduce Titan, a service-oriented framework supporting design, development, deployment, and execution of activity-aware Pervasive Apps. With Titan, mobile devices inquire surrounding nodes about available services. Internet-based application repositories compose applications based on available services as a service graph. The mobile device maps the service graph to Titan Nodes. The execution of the service graph is distributed and can be remapped at run time upon changing resource availability. The framework is geared to streaming data processing and machine learning, which is key for activity recognition. We demonstrate Titan in a pervasive gaming application involving smart dice and a sensorized wristband. We comparatively present the implementation cost and performance and discuss how novel machine learning methodologies may enhance the flexibility of the mapping of service graphs to opportunistically available nodes. H H I J J J K M
Pervasive applications need to run on collections of smart objects whose number and type are unkn... more Pervasive applications need to run on collections of smart objects whose number and type are unknown at design time and can change dynamically at runtime. This article describes the service-oriented Titan framework, where mobile phones act as mediators between dynamic and open-ended smart surroundings and Internet-based servers providing applications. Applications are composed at runtime to execute in a distributed fashion on smart objects surrounding the user. Scarce processing resources on the smart objects are efficiently managed by quick application reconfiguration. We describe the mechanisms of Titan on the example of a pervasive dice game and discuss Titan's potential as a rapid-prototyping framework for pervasive applications.
Wearable computers aim to empower people by providing relevant information at appropriate time. T... more Wearable computers aim to empower people by providing relevant information at appropriate time. This context-based information delivery helps to perform intricate, tedious or critical tasks and improves productivity, decreases error rates, and thus results in a reduction of labor cost. To evaluate the usability of wearable computing in a work environment, we have chosen a car production scenario in which new employees are trained for mechanical assembly tasks. In this paper we describe the implementation of an activity tracking system that allows to detect the individual steps of assembling a front lamp into a car body. Our approach is to break down these steps of the assembly task into smaller units, so called discrete events. Body-worn and environmental sensors are employed to create these events which trigger transitions in a Finite State Machine (FSM). The FSM states represent user activities which correspond to the assembly steps.
Google, Inc. (search). ...
This book constitutes the refereed proceedings of the Third European Conference on Smart Sensing ... more This book constitutes the refereed proceedings of the Third European Conference on Smart Sensing and Context, EuroSSC 2008, held in Zurich, Switzerland, October 29-31, 2008. The 17 revised full papers presented together with one invited paper were carefully reviewed and selected from 70 submissions. The papers are organized in topical sections on smart objects, spatial and human context inference, context processing and quality of context, as well as context-aware interaction and case studies.
2011 8th Annual IEEE Communications Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks, 2011
Energy-efficiency is key to meet lifetime requirements of Wireless Sensor Networks (WSN) applicat... more Energy-efficiency is key to meet lifetime requirements of Wireless Sensor Networks (WSN) applications. Today's run-time platforms and development environments leave it to the application developer to manage power consumption. For best results, the characteristics of the individual hardware platforms must be well understood and minutely directed. An Operating System (OS) with suitable programming abstractions can micromanage power consumption of resources. We demonstrate with the Mote Runner platform how the inherent overhead of managed application code is compensated for by a platform-independent communication API together with sleep optimizations. The proposed abstractions and optimizations can be applied to other modern sensor network platforms. To quantify the effectiveness of our approach, we measured the energy efficiency of a realworld WSN application using a custom TDMA communication protocol fully implemented on both Mote Runner and TinyOS. Mote Runner's power management and sleep phase optimizations outperforms TinyOS in our test application for duty cycles below 10% on the Iris hardware.
The development of activity recognition techniques relies on the availability of datasets of gest... more The development of activity recognition techniques relies on the availability of datasets of gestures to train and validate the proposed methods. In this work we introduce and describe a new dataset for activity recognition. The dataset is made up of 8 scenarios from everyday life and includes 17 activities composed of a total of 64 gestures. Each scenario has been repeated 10 times by 2 users. All activities and gestures are labeled. 5 different sensing modalities are implemented by using body worn and environmental sensors and smart objects. The paper describes our considerations in setting up the testbed and performing the experiments to record the dataset, our experiences with recording the data and discusses possible research questions to be tackled with the dataset.
Proceedings of the 3rd International ICST Conference on Body Area Networks, 2008
In pervasive environments, Body Area Networks (BANs) are characterized by the mobility of their u... more In pervasive environments, Body Area Networks (BANs) are characterized by the mobility of their users. BANs can continuously interact with each other, thus enabling the provision of new applications and services at runtime. New complex services can be provided by composing simpler services available on neighbouring network nodes. However, since the topology of BANs is continuously changing due to users' movements, it is unfeasible to specify a-priori all possible configurations under which a given complex service can be composed. In order to address this issue, we introduce a two-layered service discovery and composition architecture, that proactively notifies a distributed service directory with changes in service availability. In order to cope with the network mobility and intermittent connectivity, our approach is to cluster nodes in the sensor network based on their connectivity patterns. We use a multi-agent state machine to recognize the availability of complex services and to provide them. Our solution is validated by a prototype implementation of our architecture, by the study of the statistical model of complex services, and by experimental evaluations.
Lecture Notes in Computer Science, 2008
Activity recognition from an on-body sensor network enables context-aware applications in wearabl... more Activity recognition from an on-body sensor network enables context-aware applications in wearable computing. A guaranteed classification accuracy is desirable while optimizing power consumption to ensure the system's wearability. In this paper, we investigate the benefits of dynamic sensor selection in order to use efficiently available energy while achieving a desired activity recognition accuracy. For this purpose we introduce and characterize an activity recognition method with an underlying run-time sensor selection scheme. The system relies on a meta-classifier that fuses the information of classifiers operating on individual sensors. Sensors are selected according to their contribution to classification accuracy as assessed during system training. We test this system by recognizing manipulative activities of assembly-line workers in a car production environment. Results show that the system's lifetime can be significantly extended while keeping high recognition accuracies. We discuss how this approach can be implemented in a dynamic sensor network by using the context-recognition framework Titan that we are developing for dynamic and heterogeneous sensor networks.
Atlantis Ambient and Pervasive Intelligence, 2011
The automatic detection of complex human activities in daily life using distributed ambient and o... more The automatic detection of complex human activities in daily life using distributed ambient and on-body sensors is still an open research challenge. A key issue is to construct scalable systems that can capture the large diversity and variety of human activities. Dynamic system reconfiguration is a possible solution to adaptively focus on the current scene and thus reduce recognition complexity. In this work, we evaluate potential energy savings and performance gains of dynamic reconfiguration in a case study using 28 sensors recording 78 activities performed within four settings. Our results show that reconfiguration improves recognition performance by up to 11.48 %, while reducing energy consumption when turning off unneeded sensors by 74.8 %. The granularity of reconfiguration trades off recognition performance for energy savings.
Advances in Smart Systems, 2012
... Collecting datasets from Ambient Intelligence Environments Piero Zappi, University of Califor... more ... Collecting datasets from Ambient Intelligence Environments Piero Zappi, University of California San Diego, USA Clemens Lombriser, ETH Zürich, Switzerland Luca Benini, University of Bologna, Italy Gerhard Tröster, ETH Zürich, Switzerland AbSTrACT ...
The development of activity recognition techniques relies on the availability of datasets of gest... more The development of activity recognition techniques relies on the availability of datasets of gestures to train and validate the proposed methods. In this work we introduce and describe a new dataset for activity recognition. The dataset is made up of 8 scenarios from everyday life and includes 17 activities composed of a total of 64 gestures. Each scenario has been repeated 10 times by 2 users. All activities and gestures are labeled. 5 different sensing modalities are implemented by using body worn and environmental sensors and smart objects. The paper describes our considerations in setting up the testbed and performing the experiments to record the dataset, our experiences with recording the data and discusses possible research questions to be tackled with the dataset.
Uploads
Papers by Clemens Lombriser