Methods for Determining if An Ambulance is Moving Away or Approaching by Its Siren Sound
| Autor: | Fu-Yi Lin, 林福義 |
|---|---|
| Rok vydání: | 2018 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 106 This study investigates automatic methods for determining if an ambulance is moving away or approaching by its siren sound. In general, road users have to politely let the ambulances go first while hearing the siren sounds. In order to help drivers avoid being unaware of the ambulance, this study proposes to recognize every 1-sec siren sound as one of the three types, namely, approaching, away, and stop. Two approaches are developed. One is based on energy analysis for the siren sound. The other is based on long short-term memory neural network with the Mel-frequency cepstral coefficients. Our experiment results show that the recognition based on energy analysis is sensitive to noise, in which the recognition accuracy is only 45.96% for the simulated -15dB noisy siren sound data, and 55.72% in real siren sound data recorded on a downtown street. The long short-term memory neural network performs better than the energy analysis approach, which achieves 82.21% and 97.04% recognition accuracies respectively for the simulated and real siren sound data. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|