Risk Assessment and Safety-as-a-Service Provisioning in IoV Networks: A Roadmap

This paper presents a framework for a Safety recommender system that offers Safety-as-a-Service (Safe-aaS) to on-road vehicles in IoT-based transportation systems. The proposed safety recommendation system integrates several phases, including preparing sensed data, self-supervised learning from the data, and predicting personalized safety plans. The customized safety plans, generated by analyzing on-road decision parameters such as dynamic road conditions, weather patterns, and driving assistance parameters, are suggested to registered users in real time through the Safe-aaS platform tailored to their specific requirements.In a Safe-aaS enabled road transportation system, sensor nodes produce large volumes of data in a dynamic environment. Efficient analysis and storage of sensed data in edge devices poses a major challenge for service providers. Conversely, generating safety-related decisions with lower processing time ensures driver safety. The Safe-aaS-based recommender system provides safety-related decisions with low latency and processing delay. In the proposed recommendation system, a safety score is computed from the pre-processed sensed data across various parameters, followed by a user-specific similarity score and a corresponding risk score. We apply a fuzzy rule-based learning technique to predict optimized safety plans. To showcase the real-time scenario of our framework, we develop a user-friendly application interface at the primary stage. Extensive analysis shows that our system effectively detects on-road risks and adapts safety recommendations in dynamic weather and road conditions.