Establishing an AI-Driven Approach for Soundscape Assessment and Improvement in High-Density Street Design

In high-density cities, urban streets serve as vital public open spaces, fulfilling multiple functions such as transportation or social activities. However, achieving high-quality soundscape in such spaces faces challenging due to the crowds in confined spaces, which often leads to serious acoustic environmental problem. Despite the importance of this problem, soundscape considerations have rarely been integrated into the design of high-density streets as lack of targeted research. This study aims to explore the factors influencing soundscapes in high-density streets, using Shenzhen as a case study. Through in-situ investigations of 36 typical high-density streets, a critical threshold of 72 dB was identified. When sound levels exceed 72 dB, reducing the overall sound level is essential for improving the soundscape. Conversely, when sound levels are below 72 dB, factors related to visual elements and sound sources play a more significant role in soundscape optimization. Based on these findings, an AI model was developed to assess soundscape quality in high-density streets and to optimize design schemes effectively.