Related Experiment Video
Updated: Aug 13, 2026
06:37
Artificial Intelligence-Based System for Detecting Attention Levels in Students
Published on: December 15, 2023
Design of an Adaptive Model for Intelligent Car Seats Based on Human Pose Recognition
Yunpeng Bai1,2,3, Min Zhao1,2,3, Wanming Zhong1,2,3
1School of Mechatronic Engineering, Northwestern Polytechnical University, Xi'an, China.
Annals of the New York Academy of Sciences
|August 12, 2026
Summary
This study introduces an intelligent seat adjustment system using human pose recognition to enhance driver comfort and safety in electric vehicles (EVs). The model accurately estimates posture, leading to significant improvements in subjective comfort and reduced muscle strain.
Area of Science:
- Human-Computer Interaction
- Automotive Engineering
- Biomechanical Engineering
Background:
- Current electric vehicle (EV) seats lack dynamic adaptability for driver comfort and safety.
- Existing systems struggle with posture adaptation and recognizing occluded joints.
Purpose of the Study:
- To develop an intelligent seat adaptive adjustment model based on human pose recognition.
- To improve posture adaptation and recognition accuracy for occluded joints in vehicle occupants.
Main Methods:
- Utilized Kinect V2 for driver posture data collection.
- Employed a cascaded pose analysis network (PSN) with attention mechanisms for feature extraction and joint recognition.
- Integrated biomechanical and vehicle constraints with analytic hierarchy process (AHP) for seat parameter optimization.
Main Results:
- Achieved high accuracy in recognizing 16 joint points (0.87 average accuracy, 0.92 PCK@0.2, 3.2 mm MPJPE).
- Demonstrated significant improvements in simulated driving: 30% increase in subjective comfort, 39.5% decrease in lumbar muscle activity.
- Posture estimation and seat parameter generation completed within 0.3 seconds.
Conclusions:
- The proposed intelligent seat model shows preliminary feasibility for enhancing driver comfort and posture in simulated driving.
- Further validation is required for practical safety in real-world driving conditions.
