iDog: An Intelligent Guide Dog Harness for Visually Impaired Pedestrians Based on Artificial Intelligence and Edge Computing

Generally, intelligent walking canes are limited in guiding visually impaired pedestrians, which can be inconvenient. Especially in complex traffic environments, visually impaired pedestrians are less willing to use assistive devices and generally prefer guide dogs, which provide companionship and help individuals feel safe. To address this issue, an intelligent guide dog harness named iDog, which adopts artificial intelligence (AI) and edge computing technology, is proposed to improve the safety of visually impaired pedestrians. The proposed smart device is a harness for dogs that does not affect the guide dog's performance in guiding tasks. An RGB-D image sensor is used to obtain information about the surrounding environment, which is analyzed through an edge computing module and provided to visually impaired pedestrians through voice alerts. This ensures the safety of visually impaired pedestrians. The experimental results showed that the proposed iDog system can successfully recognize obstacles in traffic environments, including fast-moving obstacles and pedestrian signs, and issue voice commands when pedestrians need to cross the road. When obstacles (people, motorcycles, and cars) are identified, information about the upcoming obstacles is output through voice alerts. In addition, the proposed iDog system includes a light sensor module that detects insufficient external ambient light. When low light conditions are identified, the matrix LED light module is activated to alert other pedestrians that a visually impaired pedestrian is walking on the road.