Autonomous Driving Systems (ADS) rely on extensive testing to ensure safety and reliability, yet road scenario datasets often contain redundant cases that slow down the testing process without improving fault detection. We present a novel test prioritization approach that reduces redundancy while preserving geometric and behavioral diversity. Road scenarios are segmented into representative sections, which are compared using similarity scores based on dynamic time warping and enriched with dynamic features of the ADS driving behavior. These features guide clustering to identify groups of similar scenarios, from which representative cases are selected to guarantee coverage. Finally, we introduce a prioritization mechanism that ranks roads based on geometric complexity, driving difficulty, and historical failures, ensuring that the most critical and challenging tests are executed first. We evaluate our approach on the OPENCAT dataset and the Udacity self-driving car simulator using different ADS from the literature, showing that it substantially reduces redundancy while preserving scenario diversity and improving test efficiency.