This Project's

Research Team

Co-ordinators

Researchers

Students

External Collaborators

  • Dr Claire Guest
  • Mr Rob Harris
  • Prof. Daniel Mills

Supporting the Work of Medical Detection Dogs

Theme: Animal-Computer Interaction
Website: http://Medical Detection Dogs
Funding Agency: Medical Detection Dogs
Period: January 2013 -

This project aims to develop technology to support medical detection canines in their tasks.

Project 1. Enabling cancer detection dogs to express degrees of confidence during detection

Cancer detection dogs are trained to recognise volatiles of cancer cells in biological samples. In current practice they express a positive or negative assessment by demonstrating stereotypical behaviours (e.g. sitting down in front of or staring at a sample). However, any variations in their levels of confidence (e.g. maybe yes, maybe not, maybe), based on contamination levels, are left to the interpretation of subtleties in their body language, which can be difficult to interpret and which are highly individual. Using a combination of sensor and interactive technology, we are designing a canine interface for enabling the dogs to explicitly express degrees of confidence about a detection thus disambiguating their communicative intentions.

Project 2. Empowering medical alert dogs to respond in emergencies

Charlotte Robinson’s PhD investigates technologies to assist working canines such as medical alert dogs, but designing a canine alarm system that the dogs can use to summon help for their assisted humans. Diabetic alert dogs are trained to alert when they detect low blood sugar, up to 30 minutes before their human will notice physical signs, potentially saving their life and improving their overall quality of life. The project explores how tokens already used by the dogs to alert can be technologically enhanced to give canine users additional functionality, such as the ability to remotely alert family, friends, neighbours or emergency services, including the possibility of escalating an alert depending on the severity of an emergency.