Modeling vehicle interior noise exposure dose on freeways: Considering weaving segment designs and engine operation

Area of Science:

• Engineering
• Environmental Health
• Acoustics

Background:

• Vehicle interior noise, particularly at frequencies that may impair hearing (500 Hz and 800 Hz), is a concern for freeway commuters.
• Current noise evaluations often occur in labs, potentially underestimating real-world impacts from dynamic traffic and road conditions.
• Vehicle maneuvering, especially on complex freeway weaving segments, significantly influences interior noise levels.

Purpose of the Study:

• To explore the risk of interior noise exposure for freeway commuters on weaving segments.
• To enhance interior noise estimation by developing a decision tree learning-based noise exposure dose (NED) model.
• To incorporate weaving segment designs and engine operation into noise exposure models.

Main Methods:

• On-road driving tests were conducted with 12 subjects, collecting data via On-board Diagnosis (OBD) II, a smartphone roughness app, and a digital sound meter.
• Eleven variables, including weaving segment characteristics, were used as predictors in the model.
• The importance of predictors was assessed using out-of-bag-permuted predictor delta errors.

Main Results:

• The overall risk of hearing impairment for freeway commuters was found to be acceptable.
• Interior noise levels were most sensitive to pavement roughness, influenced by freeway configuration and traffic.
• The developed NED model demonstrated high predictive power (R = 0.93, NRMSE < 6.7%).

Conclusions:

• The study quantifies interior noise exposure on freeway weaving segments, informing commuters about potential risks.
• The constructed bagged decision tree model significantly improves interior noise estimation for road engineers and vehicle manufacturers.
• Factors like vehicle maneuvering, engine operation, pavement roughness, and weaving segment configuration are crucial for accurate noise modeling.