Dr Samuel Davey (DSTO)

Abstract: Track-Before-Detect Techniques

Modern sensor systems collect digital data and apply a sequence of signal processing algorithms to produce a higher level description of the scene. In many cases, this description consists of an estimate of the number of objects of interest (say people in a video sequence) and information about their position and velocity in space. These estimates are referred to as tracks and the objects are often called targets. Traditionally, the last two links in the signal processing chain are a single-frame detector and a tracker. The aim of the detector is to report a single point in the frame for each target, and the aim of the tracker is to associate together reports from common targets and estimate the parameters of those targets.

The traditional detect-then-track method works well for sensors with low resolution and when the target signal-to-noise-ratio (SNR) is high. However, for targets with a lower SNR, the detector must operate with a low threshold and consequently will report numerous false alarms. The tracker will tolerate these to a degree but eventually becomes swamped when the false alarm rate is too high. Track-before-detect (TkBD) is a different paradigm where the single-frame detector is removed from the processing chain and the tracker uses sensor imagery directly as its measurement signal. TkBD has been shown to provide significant improvement in detection sensitivity over the conventional detect-then-track approach, usually at the cost of increase computation cost.

There are many different methods within the TkBD paradigm. This tutorial reviews the most important methods with illustrative examples and real applications. The material is suitable for a graduate student but provides sufficient detail and reference material to guide a tracking practitioner in the application to real problems.

Bio: Dr Samuel Davey received the Bachelor of Engineering, Master of Mathematical Science and Ph.D. degrees from the University of Adelaide, Australia, in 1996, 1999, and 2003, respectively. Since 1995 he has worked for the Defence Science and Technology Organisation, Australia, in the areas of target tracking and multi-sensor fusion. He is also a Visiting Research Fellow at the University of Adelaide.

Dr Davey has published numerous papers on target tracking with particular emphasis on track initiation, clutter modeling, and track-before-detect. His paper on non-Gaussian track-before-detect won best paper at Fusion 2011 and he is the primary author of the track-before-detect section in the upcoming text Integrated Tracking, Classification, and Sensor Management: Theory and Applications.

[TOP]

Salil Kanhere (University of New South Wales)

Topic: "Participatory Sensing: CrowdSourcing Data from Mobile Phones in Urban Spaces".

Abstract: The recent wave of sensor-rich, Internet-enabled, smart mobile devices such as the Apple iPhone has opened the door for a novel paradigm for monitoring the urban landscape known as participatory sensing. Using this paradigm, ordinary citizens can collect multi-modal data streams from the surrounding environment using their mobile devices and share the same using existing communication infrastructure (e.g., 3G service or WiFi access points). Mobile phones, though not built specifically for sensing, can in fact be used as sensors. The mobile phone camera can be used as a video and image sensor. The microphone on the mobile phone, when not being used for voice conversations, can be doubled up as an acoustic sensor. The embedded GPS receiver can provide location information. Further, other embedded sensors such as gyroscopes, accelerometers and proximity sensors can collectively be used to estimate useful contextual information (e.g., is the user walking or travelling on a bicycle). The data contributed from multiple participants can be combined to build a spatiotemporal view of the phenomenon of interest and also to extract important community statistics. Given the ubiquity of mobile phones and the high density of people in metropolitan areas, participatory sensing can achieve an unprecedented level of coverage in both space and time for observing events of interest in urban spaces. Several exciting participatory sensing applications have emerged in recent years. For example, GPS traces uploaded by drivers and passengers can be used to generate real-time traffic statistics. Similarly, street-level audio samples collected by pedestrians can be aggregated to create a citywide noise map. This revolutionary paradigm is also being used to collect and share data about air pollution, cyclist experiences, diet and pricing information of consumer goods.

This tutorial aims to provide a comprehensive overview of this new and topical research area to the participants of ISSNIP 2011. Specifically, the tutorial will cover the following topics - * Overview and motivation of the new paradigm of participatory sensing * A summary of new exciting applications * A detailed discussion on the key research challenges in participatory sensing including: (i) participatory sensing architectures (ii) incentive mechanisms, (iii) trustworthiness and privacy, (iv) dealing with insufficient samples, (v) activity and context inference (vi) system design issues (vii) experimentation and large-scale data collection. * Conclusions and future directions.

This tutorial is open to researchers, academics, students and practitioners working in wireless communication and mobile computing research. It does not assume that the attendees require any prior knowledge other than basics of computer networks.

Bio: Dr. Salil Kanhere received his M.S. and Ph.D. degrees, both in Electrical Engineering from Drexel University, Philadelphia in 2001 and 2003, respectively. He is currently a Senior Lecturer in the School of Computer Science and Engineering at the University of New South Wales in Sydney, Australia. His research interests include sensor networks, mobile networks and vehicular communication. He has published over 80 peer-reviewed articles on these research topics. He currently serves as the Area Editor for the ICST Journal on Ubiquitous Environments and the European Transactions on Telecommunications. He has also served on the organizing committee and program committee of a number of IEEE and ACM international conferences. Salil is a Senior Member of the IEEE and ACM.

[TOP]

Workshop on Water Quality Monitoring. Friday 9/11/2011 11.15am - 3.30pm

Panel of experts
Prof Chris Saint, Director; SA Water Centre for Water Management and Reuse, University of South Australia
Dr Michele Akeroyd; Program Manager - Drinking Water, Water Quality Research Australia
Dr Nick Crosbie; Acting Team Leader, Drinking Water Quality Planning, Melbourne Water, and Convenor, AWA Water Quality and Analysis Specialist Network
Prof Richard Stuetz; Co-Director, UNSW Water Research Centre, University of New South Wales, School of Civil and Environmental Engineering.
Prof John Bridgeman; Reader in Environmental Engineering in the School of Civil Engineering, University of Birmingham, UK

The accompanying workshop session will consider such issues as the robustness and accuracy of currently available sensors and the issues surrounding operator analysis of data and the provision of timely and easily interpretable data. The workshop is structured to address the current challenges:
• There is a growing need for the development of on-line sensors for our water, wastewater and reclaimed water areas of the industry.
• Changing environmental conditions mean that the quality of our source waters are showing increased variability and early detection of changes in quality and impacts on the treatment process are vital.
• In the wastewater area there is constant pressure to improve the efficacy of our processes, especially with regard to the removal of nutrients and other key contaminants. Improved on-line monitoring will permit better optimisation of the treatment train. Another factor that can adversely impact on the wastewater treatment process are shock loads of contaminants such as trade wastes entering the system unexpectedly, on-line monitoring has the potential to give a rapid indication of the changing quality of incoming wastewater and will allow improved management methods to be adopted at an early stage in the process.
• With regards to re-claimed water such as stormwater, the raw product is known to be highly variable and many schemes are designed to allow the separation of ‘first flush’ water from the main body of the treatment system. However, such procedures are far from optimised and by the time laboratory analysis is carried out, and data confirmed, stormwater that is potentially of good quality and high value can be wasted.

The workshop will have three sections:
1. An overview, including two introductory talks presented by Dr Nick Crosbie, Melbourne Water and AWA Water Quality and Analysis Specialist Network convenor and Prof Chris Saint, CWMR, UniSA.
2. Current applications of sensors for WQ Monitoring - showcasing current industrial based R&D projects will include several short presentations by industry researchers.
3. A panel discussion, which will be chaired by a group of experts. The format will be started from:
a) Water industry point of view, current industry practice and what improvements are required, water industry representatives will guide us on the current issues.
b) What is available and what can be adapted to our needs? Representatives from various service providers including instrument suppliers will join the discussion to show what ready solutions are available and
c) Research Direction will include University researchers to show how their research can provide future solutions.

More details see flyer.

[TOP]