Efficient Non-Compression Auto-Encoder for Driving Noise-based Road Surface Anomaly Detection
This work addresses real-time road condition monitoring for vehicle safety, presenting an incremental improvement in efficiency and performance for a domain-specific application.
The paper tackles road surface anomaly detection using driving noise to prevent accidents, proposing a non-compression auto-encoder that reduces computational cost by up to 25 times and improves anomaly detection performance by up to 7.72%.
Wet weather makes water film over the road and that film causes lower friction between tire and road surface. When a vehicle passes the low-friction road, the accident can occur up to 35% higher frequency than a normal condition road. In order to prevent accidents as above, identifying the road condition in real-time is essential. Thus, we propose a convolutional auto-encoder-based anomaly detection model for taking both less computational resources and achieving higher anomaly detection performance. The proposed model adopts a non-compression method rather than a conventional bottleneck structured auto-encoder. As a result, the computational cost of the neural network is reduced up to 1 over 25 compared to the conventional models and the anomaly detection performance is improved by up to 7.72%. Thus, we conclude the proposed model as a cutting-edge algorithm for real-time anomaly detection.