Description

Problem: 

Road traffic noise is a major environmental problem, especially in urban areas, and the researchers want to assess how well noise levels can be reduced to improve the quality of life for residents. The biggest source of this noise is the engine sound of cars. Engine noise inside and outside the car has decreased with the spread of electric cars. The biggest source of noise is the noise between the tire and the road. Furthermore, a significant proportion of traffic noise comes from it. Many studies have focused on improving this and the sound quality of tire/road noise. The focus of this article is on data-driven solutions. There are many techniques that can be used for this purpose, but the most popular today is the use of artificial intelligence in monitoring. Therefore, the emphasis is on data, signal processing, feature extraction, feature selection and artificial intelligence and signal processing algorithms that have been used in this area and that can be used in the future. Unlike literature, algorithms that have been used in many fields in recent years but have not been used much in this field have been preferred. The use of Generative adversarial networks (GANs), AutoEncoder, LSTM, GRU (Gated Recurrent Unit) algorithms will be investigated in this field.

Goal: The aim of this study is to estimate road-tire noise using signal processing and artificial intelligence algorithms using data received from time-dependent sensors such as temperature, vibration and sound. First of all, meaningful data will be selected, and its size will be reduced by signal processing, feature extraction and selection methods of sensor data. Then, a prediction model will be developed with artificial intelligence algorithms. Studies on transfer learning, hyperparameters selection, and cross validation will be carried out to increase the performance of the algorithms. In the algorithm section, research will be conducted on their hybrid use. On example approach that could be explored is Autoencoder-LSTM.

Required Skills and Knowledge:

• Programming skills (preferably in Phyton or C++)
• Basic knowledge of data analysis, signal processing libraries (PyTorch, Cuda, Pandas, NumPy, ScikitLearn, TensorFlow, Keras).
• To have basic knowledge about signal processing and time series data. Or the willingness to learn these things.