🤖 AI Summary
To address weak model generalization and insufficient control stability in micro autonomous cars (MACs) operating within simulation environments, this paper proposes an end-to-end driving framework leveraging RGB-D data and a 3D convolutional neural network (3D CNN). Unlike conventional RNN-based approaches, the method directly learns driving policies from spatiotemporally continuous RGB-D image sequences, thereby enhancing modeling of dynamic track scenes through intrinsic spatiotemporal feature extraction. Experiments conducted on two simulated tracks with distinct complexity levels demonstrate that the proposed approach achieves high lap completion rates and consistent lap times. It significantly outperforms RNN baselines in task success rate and trajectory smoothness. These results validate the efficacy of 3D CNNs in improving control accuracy and environmental adaptability for resource-constrained, miniature autonomous driving systems.
📝 Abstract
Autonomous driving applications have become increasingly relevant in the automotive industry due to their potential to enhance vehicle safety, efficiency, and user experience, thereby meeting the growing demand for sophisticated driving assistance features. However, the development of reliable and trustworthy autonomous systems poses challenges such as high complexity, prolonged training periods, and intrinsic levels of uncertainty. Mini Autonomous Cars (MACs) are used as a practical testbed, enabling validation of autonomous control methodologies on small-scale setups. This simplified and cost-effective environment facilitates rapid evaluation and comparison of machine learning models, which is particularly useful for algorithms requiring online training. To address these challenges, this work presents a methodology based on RGB-D information and three-dimensional convolutional neural networks (3D CNNs) for MAC autonomous driving in simulated environments. We evaluate the proposed approach against recurrent neural networks (RNNs), with architectures trained and tested on two simulated tracks with distinct environmental features. Performance was assessed using task completion success, lap-time metrics, and driving consistency. Results highlight how architectural modifications and track complexity influence the models' generalization capability and vehicle control performance. The proposed 3D CNN demonstrated promising results when compared with RNNs.