HCC: Small: FIDO - Facilitating Interaction for Dogs with Occupations: Wearable computing for two-way communication with assistance dogs

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Assistance dogs have improved the lives of thousands of people with disabilities. Guide dogs, service dogs, and hearing dogs can provide independence and significantly enhance quality of life. However, communication between human and canine partners is currently limited. Handlers give commands, and dogs respond with behaviors, which can sometimes have ambiguous meanings to the handlers. For example, a guide dog stops while walking with his handler, but the dog has no definitive way to tell the handler there is a new barrier on a familiar path. A hearing dog alerts his handler, but can't easily report that the fire alarm has sounded and they need to evacuate. A service dog may need to receive a "lie down" command silently in a dark, crowded theater, or from a distance, which would preclude verbal or hand signals. The PI's goal in this research is to explore fundamental aspects of wearable technologies to support two-way communication between assistance dogs and their handlers. To this end, the PI will investigate on-body interfaces for dogs in the form of electronic textiles and computers integrated into assistance dog clothing, such as vests and harnesses. Studies will assess the abilities of dogs to interact with affordances worn on their bodies, and determine what stimuli dogs can sense and comprehend from wearable technology. The work will be divided into three thrusts: Dog-to-handler communication (to determine to what extent dogs can activate affordances on their harnesses, vests, coats, or collars, such as "touch points" which can be activated by a dog's nose touch, and "pull points" which can be activated by tugging); Handler-to-dog communication (to investigate the sensory capabilities of dogs to respond to simple stimuli incorporated into a dog vest or harness, such as small vibrating motors); Handler feedback and control (the team will implement a head-mounted visual and/or auditory display to allow handlers to receive input from their dogs, as well as control interfaces that can be integrated into a harness handle, leash, or a piece of clothing worn by the handler). Design and testing of the on-body interfaces for the dogs will be carried out using an iterative approach, and the final designs will be validated with three assistance-trained dogs in laboratory and real-world environments. The PI will also stress-test the designs with dogs at speed on an obstacle course. Broader Impacts: Project outcomes will extend the state of the art in wearable interfaces to animals, and will contribute pioneering work to the nascent fields of Animal Computer Interfaces and Inter-species Interaction, as well as discovering more about the physical and cognitive abilities of dogs. Adapting usability analysis and design techniques from the human realm to apply to animals will expand the body of knowledge in Interactive Computing. In addition to the direct impact of improving the quality of life for people with disabilities, the technologies developed in this project could have application to other working dog teams (e.g., in military or police canine units, and in search and rescue), and they may ultimately allow all pet owners to better communicate with and train their dogs.