| Proj No. | A3052-251 |
| Title | Residential dataset analysis |
| Summary | In a groundbreaking initiative aimed at enhancing urban living conditions, our team embarked on a comprehensive project to gather a vast array of in-situ residential noise data emanating from various structural and air-borne sources within a Housing and Development Board (HDB) building. The scope of this endeavor was not only to map the auditory landscape of high-density living areas but also to pave the way for advanced noise classification solutions through machine learning. Students participating in this project are tasked with a multifaceted analysis of the collected datasets. The initial phase involves rigorous tagging of the noise samples, categorizing them according to their sources and characteristics, such as frequency, amplitude, and duration. This step is crucial for the creation of a structured dataset that can effectively train machine learning models. Following the tagging process, the project demands sophisticated signal preprocessing techniques. Students will engage in filtering, normalization, and feature extraction, among other methods, to refine the audio data. This preprocessing is essential to eliminate background noise and enhance the quality of the signals, thereby increasing the accuracy of the subsequent analysis. The core analytical phase requires students to apply advanced signal processing and machine learning algorithms to classify the noise types accurately. This involves training deep neural networks (DNNs) on the preprocessed datasets, tuning parameters, and testing the models to ensure high levels of precision and reliability in noise classification. Through iterative refinement, the project aims to develop a robust model capable of identifying and characterizing different types of noise with minimal error. Once the models are developed and validated, the final objective is to make the datasets publicly available on DR-NTU, the digital repository of Nanyang Technological University. This open-source initiative will not only contribute valuable resources to the global research community but also foster further innovation in the field of urban noise analysis and mitigation. Engaging in this project offers students an unparalleled opportunity to develop and hone their skills in big data analysis, signal processing, and machine learning. They will navigate through the challenges of handling large-scale datasets, apply theoretical knowledge to real-world problems, and contribute to the development of solutions that have the potential to improve the quality of life in densely populated urban areas. Through this hands-on experience, students will emerge as proficient data scientists equipped with a deep understanding of the complexities involved in environmental noise analysis and the application of artificial intelligence in solving such issues. |
| Supervisor | Prof Gan Woon Seng (Loc:S2 > S2 B2B > S2 B2B 68, Ext: +65 67904538) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | Digital Signal Processing Lab (Loc: S2-B4a-03) |
| Single/Group: | Single |
| Area: | Digital Media Processing and Computer Engineering |
| ISP/RI/SMP/SCP?: |