PhD Position Context Recognition in Opportunistic Sensor Networks

Background

State of the art systems assume statically defined sensor configuration, where the location of the sensor and its characteristics is known a-priori and does not change.

In a real-world scenario this is not the case: sensor location on body may change, sensor characteristics may degrade over time, sensors may be added or removed in an instrumented environments. The devices the user carry with him change depending on the activities.

The new European project OPPORTUNITY will develop new methods for context and activity recognition in opportunistic sensors configurations.

EU project OPPORTUNITY

OPPORTUNITY picks up on the very essential methodological underpinnings of any Ambient Intelligence (AmI) scenario: recognizing (and understanding) context and activity.

Methodologies are missing to design context-aware systems: (1) working over long periods of time despite changes in sensing infrastructure (sensor failures, degradation); (2) providing the freedom to users to change wearable device placement; (3) that can be deployed without user-specific training. This limits the real-world deployment of AmI systems.

We develop opportunistic systems that recognize complex activities/contexts despite the absence of static assumptions about sensor availability and characteristics. They are based on goal-oriented sensor assemblies spontaneously arising and self-organizing to achieve a common activity/context recognition goal. They are embodied and situated, relying on self-supervised learning to achieve autonomous operation. They makes best use of the available resources, and keep working despite-or improves thanks to-changes in the sensing environment. Changes include e.g. placement, modality, sensor parameters and can occur at runtime.

Four groups contribute to this goal. They develop: (1) intermediate features that reduce the impact of sensor parameter variability and isolate the recognition chain from sensor specificities; (2) classifier and classifier fusion methods suited for opportunistic systems, capable of incorporating new knowledge online, monitoring their own performance, and dynamically selecting most appropriate information sources; (3) unsupervised dynamic adaptation and autonomous evolution principles to cope with short term changes and long term trends in sensor infrastructure, (4) goal-oriented cooperative sensor ensembles to opportunistically collect data about the user and his environment in a scalable way.

The methods are demonstrated in complex opportunistic activity recognition scenarios, and on robust opportunistic EEG-based BCI systems.

Job description

We offer a PhD position within the framework of the new European Research Project OPPORTUNITY.
OPPORTUNITY groups 4 high-profile European universities and research institutes. They will collaborate over the next 3 years (2009-2011) to develop systems capable of activity recognition from sensors opportunistically discovered on the user and in his neighborhood. This includes hardware, software and algorithmic innovations that will be combined in a number of technology demonstrators.

In this position you will be responsible for one of the project's work package. You will closely collaborate with the project's partners throughout the duration of the project.

Your work environment will be multinational, both in Zürich and with project partners within Europe, with frequent travels to the partner's location.

Within this project, your research topics will include (but are not limited to):

• Context recognition in sensor networks.
  
• Unsupervised dynamic adaptation. Methods will be investigated and developed to achieve a dynamically-defined multi-parametric performance goal (e.g. accuracy, energy use) given an opportunistically discovered set of sensors and confidence and signal quality metrics. The methods will focus on the dynamic adaptation of the activity recognition chain (e.g. through  dynamic resource selection, feature, and classifier adaptation) on the basis of a model of the system performance.
  
• Autonomous evolution in open-ended environments. New sensors, that were not available when the system was trained, can be discovered at run-time. Methods will be developed to take advantage of additional resources without the need for user intervention or system re-training. First approaches will focus on self-supervised learning - where a part of the system supervises the training of the new part of the system. Our end objective is to achieve autonomously evolving context-aware systems capable to "grow" and use new resources as they are discovered, in open-ended environment.
• Inclusion of minimal user feedback. Methods will be devised so that minimally intrusive user feedback can be used by the system to enhance activity recognition performance (e.g. sporadic annotation of an activity by the user). Constraints include maximizing information gain and minimizing user disturbance.
  

Starting date: ASAP

Requirements

The candidate has a diploma, MSc, or equivalent in electrical engineering, micro-engineering, computer science or mathematics.

He has strong interests in mobile computing systems, machine learning/pattern recognition, signal processing, adaptive and learning systems, and in the combination of theoretical and experimental research.

Fluent spoken and written English is mandatory.

Contact and application

For further information about the Euopean project Opportunity and your contribution within this project, contact Dr. Daniel Roggen.

If you are interested and believe that you qualify, please send your application to Prof. Gerhard Tröster. Include:

• Curriculum Vitae ith the names and contact details of at least 2 references
• a list of exams and grades obtained
• a cover letter explaining how your skills and research interests fit the project