An evaluation model for automobile intelligent cockpit comfort based on improved combination weighting-cloud model

Area of Science:

• Automotive Engineering
• Human-Computer Interaction
• Artificial Intelligence

Background:

• Assessing automobile intelligent cockpit comfort is crucial for user experience.
• Existing evaluation methods may not fully capture the complex, fuzzy nature of comfort factors.
• Intelligent cockpits present new challenges for comfort evaluation due to integrated systems.

Purpose of the Study:

• To develop an advanced comfort evaluation model for automobile intelligent cockpits.
• To integrate subjective and objective weighting methods for a more robust assessment.
• To enhance comfort evaluation accuracy by addressing index system fuzziness and randomness.

Main Methods:

• Established a comfort evaluation system with 4 first-class and 15 second-class indexes (e.g., noise, light, thermal, HCI).
• Combined subjective and objective weights using improved Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) via Game Theory.
• Utilized a cloud model with floating cloud algorithms and improved similarity calculations for comprehensive evaluation.

Main Results:

• The developed model successfully determined first-class, second-class, and comprehensive evaluation cloud parameters.
• Improved similarity calculation methods (ECM, MCM) optimized evaluation results.
• Verification using a 2021 Audi intelligent car confirmed the model's correctness and rationality.

Conclusions:

• The improved combination weighting-cloud model provides a more accurate reflection of automobile intelligent cockpit comfort.
• The model effectively handles the inherent fuzziness and randomness in comfort evaluation.
• This approach offers a valuable tool for automotive manufacturers to enhance cockpit design.