ESR1. Complex activity modelling and recognition within smart environments.

Phd enrollment: Yes.

Start Date: June - September 2016.

Duration: 36 Months.

Host Organisation: School of Computer Science and Informatics, De Montfort University, UK

Location: The Gateway, Leicester, LE1 9BH, UK


Project Title: Complex Activity Modelling and Recognition within Smart Environments

Objectives: Context inference and activity recognition play a critical role in assisted living. Extensive research has been undertaken in this area, including data-driven methods based on data mining and machine learning techniques, e.g. HMM, DBN, SVM, CRF, and knowledge-driven methods based on knowledge modelling, representation and inference, e.g. event calculus, plan recognition, ontological modelling and reasoning. Nevertheless, current research mainly focuses on simple, pre-scripted and well-designed scenarios. Many problems remain open issues, e.g. interleaved and concurrent activity recognition, which are important for behaviour analysis, health prediction and personalised assistance.

This project will develop a novel complex activity recognition system which is capable of discerning real-time sensor observations on-the-fly and dynamically recognising simple and complex (e.g. interleaved or concurrent) activities. The system will support continuous progressive activity analysis and provide context-aware assistance, in particular, for people with cognitive impairments. This will be achieved by developing 1) semantically rich sensor and activity models, 2) a real time sensor data separation method based on sensor observation semantics, 3) a multi-thread algorithm for progressive multiple atomic and complex activity recognition, and 4) a working prototype system towards a demonstrator.  

Expected Results

  • Underlying semantic sensor models and dynamic segmentation algorithms.
  • Semantic enabled recognition methods.
  • An implemented prototype for an overall assistive living solution.
  • Integration of technologies and tools into application demonstrators.

Planned Secondments:

  • 1st: ACCORD, UK, to engage with professionals and understand end users’ needs and considerations.
  • 2nd: Dundalk Institute of Technology, Ireland, to understand the CASALA living lab and the Great Northern Haven and dada sets.
  • 3rd: Universidad Politecnica De Madrid, Spain, to know living support lab, prepare and conduct benchmark evaluation.
  • 4th: Northwestern Polytechnical University, China, to get an international perspective and research exchange.

How to Apply:

The post advertisement, including eligibility requirements, Job Specs and Person Specs can be found in the Recruitment Section of the ACROSSING project website.

To apply, follow instructions in the university’s research degree online application page.

For further details or enquiries please contact the Graduate School Office via researchstudents@dmu.ac.uk or Prof Liming Chen on +44 (0)116 2078490 or email liming.chen@dmu.ac.uk.

Please quote the post reference number to which you would like to apply.