The appearance of snowbanks, ice and large slush puddles can quickly turn walking outdoors into a dangerous and more difficult environment to navigate for older adults and individuals with disabilities. This population is at high risk of experiencing falls and for having their travel slowed down by challenging weather conditions. A fall-related injury can spark a sudden downward spiral in health and increased difficulty actively traversing through the outdoor built environment can cause many to become discouraged to go outside in the winter, possibly due to a fear of experiencing a fall. These feelings can lead to individuals limiting their physical activity outdoors in the winter resulting in decreased social interaction and isolation.
Despite the magnitude of the problem, there has been little research focused on understanding how the design of our outdoor built environment can be improved to meet the needs of high-risk pedestrians. The objectives of this project are to:
Develop a machine learning system for tracking the walking speed and trajectory of pedestrians using birds-eye-view video recordings of street crossings and,
Identify differences in pedestrian behaviour for individuals who fall into the 5th percentile for walking speed and below compared to pedestrians in the top two quartiles for walking speed in challenging weather conditions.
This work will provide a new tool that will allow real-world testing will be conducted using machine learning approaches, to efficiently assess the speed and movement of large numbers of pedestrians quickly. The findings of this work will be used to improve the design and maintenance of street crossings to reduce the negative effects of challenging weather conditions on individuals with disabilities.
This research is ongoing and results are not yet available.