Active Data Acquisition in Autonomous Driving Simulation
This addresses the high costs and inefficiencies in data acquisition for autonomous driving systems, though it appears incremental as it builds on existing active learning concepts.
This paper tackles the problem of redundant and costly data collection for autonomous driving by proposing an active data-collecting strategy, which reduces labeling costs and dataset size while improving dataset quality and system performance.
Autonomous driving algorithms rely heavily on learning-based models, which require large datasets for training. However, there is often a large amount of redundant information in these datasets, while collecting and processing these datasets can be time-consuming and expensive. To address this issue, this paper proposes the concept of an active data-collecting strategy. For high-quality data, increasing the collection density can improve the overall quality of the dataset, ultimately achieving similar or even better results than the original dataset with lower labeling costs and smaller dataset sizes. In this paper, we design experiments to verify the quality of the collected dataset and to demonstrate this strategy can significantly reduce labeling costs and dataset size while improving the overall quality of the dataset, leading to better performance of autonomous driving systems. The source code implementing the proposed approach is publicly available on https://github.com/Th1nkMore/carla_dataset_tools.