ASTRO-K: Finding Top-k Sufficiently Distinct Indoor-Outdoor Paths

CAPRIO is an indoor-outdoor pedestrian path rec-ommendation system that optimizes for shortest distance. Its path-finding algorithm, ASTRO, takes into account a set of user-provided congestion constraints and as such can recommend paths that can reduce the risk of COVID-19 exposure. In this paper, we extend ASTRO to consider the changes on congestion when providing path recommendations for overlapping requests. Our new algorithm, called ASTRO-K, can provide K alternative paths that satisfy the congestion constraints of all the path requests within a short time-window. Our experimental eval-uation is conducted using two real-world datasets and shows that ASTRO-K can reduce the total average congestion of the recommended paths up to 4.5X with the trade-off of up to 7% increased total path time.