Research

School of Science and Technology Computing About Us Research

Research Areas

HKMU Computing has a small staff focusing on several core research areas.


Meta-heuristic and Evolutionary Algorithms

Our focus is on the family of multi-objective optimization algorithms and its adaptation for tackling challenging problems. For example, we studied Multi-Objective Particle Swarm Optimizer (MOPSO), which is an effective swarm intelligence algorithm simulating social behavior of birds within a flock, and improved its performance.

List of projects carried out in this area:

  • Leader selection strategies for improving convergence and diversity
  • Multiple objective data clustering
  • Mechanism for balance of exploration and exploitation
  • Cluster merge-split modelling in temporal clustering
  • Student performance modelling for learning analytics

Improved convergence with adaptive inertia weight and dynamic search space

Adaptive student performance clustering


Educational Uses of Virtual Reality and Augmented Reality

Virtual Reality (VR) and Augmented Reality (AR) technologies offer an highly-configurable context for learning scenarios of which the effectiveness rests on the authenticity of the learning context.

Our research and development direction is on studying effective design principles for applying VR/AR in different subject areas and scenarios. One example is hospital ward orientation for nurse education. A VR environment ensures students can explore in a safe manner, and in particularly no real patient would be threatened. A virtual hospital ward with interaction support was developed for nurse students to have some experience with the environment before their actual practice in a hospital ward. Students could interact with objects that are relevant to their practicum.

A ward and the floor plan in the virtual hospital

Various objects that are relevant to students when they attend their practicum


Applications of Radial Basis Functions (RBFs)

The RBFs were originally devised for scattered geographical data interpolation by Hardy, who introduced a class of functions called multiquadric functions in the early 1970’s. The basic idea of the RBF interpolation is to approximate an unknown function by an interpolant at a set of N distinct data points, and the interpolant depends on the Euclidean distance from the origin. One powerful use of RBFs is to solve partial differential equations. RBFs have been successfully applied in many different problems like stress problems, fluid mechanics etc. Here we present how the famous Motz’s problem can be solved using RBFs.

The the L shape region with various governing equation and boundary conditions

A plot of the solution found by RBFs

 

Videos for Teaching and Learning

For many years, we have developed expertise in using computing technologies to enhance teaching and learning. Video has been heavily used in instructional contexts since the thriving days of distance learning education. Nowadays, the educational use of video has not limited to instructional videos (talking-heads and class recordings). A number of innovations have been attempted and studied such as video modelling, and digital storytelling.

List of projects carried out in this area:

  • Video modelling for business etiquette
  • Student generated video for meta-cognitive assessment
  • Video team project for fostering sense of belonging
  • Gesture based interaction for seamless delivery of video lectures

Gesture based interaction for seamless delivery of video lectures

Video modelling for business etiquette

Video modelling for business etiquette

Internet-of-Things (IoT)

Our work in this area includes algorithm and application development.

List of projects carried out:

  • Indoor positioning augmented with environmental landmark detection
  • Health related applications of IoT
  • Internet-of-Things based serious games
  • Campus IoT for fostering sense of belonging