CrowdSPaFE: A Crowd-Sourced Multimodal Recommendation System for Urban Route Safety

Abstract

Navigation and traffic services such as Google Maps, Bing Maps, and Apple Maps have become increasingly popular for their ability to calculate the shortest path, provide real-time traffic updates, recommend nearby points of interest, and suggest multi-modal route options based on user constraints. However, while these services offer convenience and efficiency, they may not always prioritize user safety. In response to this concern, recent research have begun to address safety issues in navigation and traffic services. To the best of our knowledge, none of these are capable of adapting to dynamic, conflicting safety features and real-time user feedback. A recent algorithm called SPaFE has been introduced to incorporate crowd-sourced and historical data, but it does not prioritize the most recent feedback or consider updated crime reports. It also does not account for distance and performs poorly in areas with insignificant or zero feedback. In light of the preceding, we introduce CrowdSPaFE, a population-based algorithm that adapts to dynamic crime reports, the most recent feedback, navigation in locations with negligible feedback, and a tradeoff between distance and safety considerations. Lastly, our empirical results demonstrate that the CrowdSPaFE algorithm outperforms the state-of-the-art.

Keywords

• Crowd-sourcing
• population based safe path algorithm
• risk minimization problem
• ant colony optimization
• graph theory