Markus Scholz

RF-Sensing of Activities from Non-Cooperative Subjects in Device-Free Recognition Systems Using Ambient and Local Signals

We consider the detection of activities from non-cooperating individuals with features obtained on the radio frequency channel. Since environmental changes impact the transmission channel between devices, the detection of this alteration can be used to classify environmental situations. We identify relevant features to detect activities of non-actively transmitting subjects. In particular, we distinguish with high accuracy an empty environment or a walking, lying, crawling or standing person, in case-studies of an active, device-free activity recognition system with software defined radios. We distinguish between two cases in which the transmitter is either under the control of the system or ambient. For activity detection the application of one-stage and two-stage classifiers is considered. Apart from the discrimination of the above activities, we can show that a detected activity can also be localized simultaneously within an area of less than 1 meter radius.

Device-free and device-bound activity recognition using radio signal strength

Background: We investigate direct use of 802.15.4 radio signal strength indication (RSSI) for human activity recognition when 1) a user carries a wireless node (device-bound) and when 2) a user moves in the wireless sensor net (WSN) without a WSN node (device-free). We investigate recognition feasibility in respect to network topology, subject and room geometry (door open, half, closed). Methods: In a 2 person office room 8 wireless nodes are installed in a 3D topology. Two subjects are outfitted with a sensor node on the hip. Acceleration and RSSI are recorded while subject performs 6 different activities or room is empty. We apply machine learning for analysis and compare our results to acceleration data. Results: 10-fold cross-validation with all nodes gives accuracies of 0.896 (device-bound), 0.894 (device-free) and 0.88 (accelerometer). Topology investigation reveals that similar accuracies may be reached with only 5 (device-bound) or 4 (device-free) selected nodes. Applying trained data from one subject to the other and vice-versa shows higher recognition difference on RSSI than on acceleration. Changing of door state has smaller effect on both systems than subject change; with least impact when door is closed. Conclusion: 802.15.4 RSSI suited for activity recognition. 3D topology is helpful in respect to type of activities. Discrimination of subjects seems possible. Practical systems must adapt no only to long-term environmental dispersion but consider typical geometric changes. Adaptable, robust recognition models must be developed.

A flexible architecture for a robust indoor navigation support device for firefighters

In harsh indoor environments like in a firefighter operation location technologies would help to reduce casualties. However, exact indoor localization is still a research topic. We aim to create a wireless sensor network based ad hoc system which builds on the existing navigational skills of firefighters. Allowing them to shape the system as it best fits the actual operation will enhance efficiency at affordable costs. Such a system could provide the advantages of a fixed infrastructure independent of the place of action. In this paper we first analyze the architectural requirements of such a system. Second, we present a corresponding three layered system design which is comprised of network, data management and data storage layer. Third, an implementation of the architecture is presented. Fourth, a prototype implementation of the system and finally, a report on the system evaluation is given. The designed architecture is a promising approach towards a robust and flexible indoor navigation support device for firefighters.

Landmarke: an ad hoc deployable ubicomp infrastructure to support indoor navigation of firefighters

Indoor navigation plays a central role for the safety of firefighters. The circumstances in which a firefighting intervention occurs represent a rather complex challenge for the design of supporting technology. In this paper, we present the results of our work designing an ad hoc ubicomp infrastructure to support navigation of firefighters working in structure fires inside the zone of danger. We take a wider approach, complementing the technical questions with the development of effective navigation practices based on technology available today. We provide an overview of the complete design process, from the theoretical and empirical underpinnings to the construction and evaluation of three iterations of the platform. We report the results of our evaluation and the implications and tensions uncovered in this process, and we discuss the challenges and implications of it for the design of ubicomp for firefighters.

Device-Free Radio-based Low Overhead Identification of Subject Classes

An increasing corpus of research focuses on inferring contexts solely through analysis of changes in surrounding wireless signals without the subject carrying a device (device-free). This paper takes device-free recognition a step further: We present WiDisc, a novel device-free RF system for distinguishing three subject classes (e.g. tall, medium, small). WiDisc models the problem as fingerprinting-based classification. To alleviate the significant location-based training overhead per subject class which is usually required, WiDisc employs 3D subject class model construction and electromagnetic simulations to generate the fingerprints with no manual training overhead. WiDisc further estimates the most relevant RF links to maximize recognition performance. Our lab evaluation with only four transceivers and three subject classes shows that the link selection module can accurately predict the two most important links, falling short only 5% of the achievable accuracy. In addition, WiDisc achieves a classification accuracy of 67% with zero training overhead vs 76% with traditional fingerprinting. Discrimination works esp. well for the medium and tall subjects but confusions for the small subject are frequent, indicating potential for further research. Still, the results highlight WiDiscs ability to trade off accuracy and training overhead and opens the door for new applications including finer-grained intrusion detection forensics, device-free parental control, personalized device-free gesture recognition, to name a few.

A model driven internet of things

The presented work proposes a model driven development approach employing flexible code generation strategies to overcome the technological gap between networked embedded objects and enterprise back-end system. We design a modeling and generator toolchain based on state of the art technologies to support efficient data exchange between hybrid resource constrained systems. We start by mapping different high level message description languages to a common meta model. From those models we generate visual pushdown automata that are the basis for our code generation and message encoding strategy. On the basis of this representation we present a practical and efficient approach to generate binary representation.

Passive, Device-Free Recognition on Your Mobile Phone: Tools, Features and a Case Study

We investigate the detection of activities and presence in the proximity of a mobile phone via the WiFi-RSSI at the phone. This is the first study to utilise RSSI in received packets at a mobile phone for the classification of activities. We discuss challenges that hinder the utilisation of WiFi PHY-layer information, recapitulate lessons learned and describe the hardware and software employed. Also, we discuss features for activity recognition (AR) based on RSSI and present two case studies. We make available our implemented tools for AR based on RSSI.

jNode: A sensor network platform that supports distributed inertial kinematic monitoring

Because of the intrinsic advantages of wireless inertial motion tracking, standalone devices that integrate inertial motion units with wireless networking capabilities have gained much interest in recent years. Several platforms, both commercially available and academic, have been proposed to balance the challenges of a small form-factor, power consumption, accuracy and processing speed. Applications include ambulatory monitoring to support healthcare, sport activity analysis, recognizing human group behaviour, navigation support for humans, robots and unmanned vehicles, but also in structural monitoring of large buildings. This paper provides an analysis of the current state-of-the-art platforms in wireless inertial motion tracking and presents a novel open-source and open-hardware hybrid tracking platform that is extensible, low-power, flexible enough to be used for both short- and long-term monitoring and based on a firmware that allows it to be easily adapted after being deployed.

Evaluation of wireless sensor technologies in a firefighting environment

In firefighter environments navigational support could help to reduce casualties. While exact indoor localization is still a research problem, an alternative may be a bread crumb based approach in which not the exact localization but the recovery of laid out wireless sensor nodes is crucial. Recovery, however, can be enabled using sensors typically employed for indoor positioning in wireless sensor networks (WSNs). Such sensors include infrared, ultrasound and radio. Only little information is available on the behavior of these sensors under the influence of a firefighter environment. In this paper first a report of the evaluation of these sensor technologies under the harsh conditions in a firefighting training facility is given. Secondly, tests considering received signal strength in respect to firefighter equipment, postures, movement patterns and antenna positions are presented. Third, two potential antenna configurations are evaluated. We show that the evaluated sensor technologies and antennas may be used to realize the envisioned navigation tool. We conclude that antenna placement is crucial and propose the front side of the helmet as optimal location for a directional antenna.

Global peer-to-peer classification in mobile ad-hoc networks: a requirements analysis

This paper examines global context classification in peer-to-peer ad-hoc mobile wireless networks (P2P-MANETs). To begin, circumstances are presented in which such systems would be required to classify a global context. These circumstances are expounded upon by presenting concrete scenarios from which a set of requirements are derived. Using these requirements, related work is evaluated for applicability, indicating no adequate solutions. Algorithmic approaches are proposed, and analysis results in a benchmark as well as bounds for distribution of processing load, memory consumption and message passing in P2P-MANETs.