An LSTM-Based Autonomous Driving Model Using Waymo Open Dataset
🤖 AI Summary
To address the bias in academic driving behavior modeling caused by proprietary, closed-loop strategies employed by autonomous driving companies, this paper introduces, for the first time, an end-to-end inverse learning framework for black-box driving policies, built upon the Waymo Open Dataset. We employ an LSTM to model multi-modal temporal sensor data (LiDAR, camera, IMU) and directly predict low-level control actions—steering angle and acceleration—using mean absolute error (MAE) as the optimization objective, thereby eliminating reliance on hand-crafted rule-based assumptions. We further propose an interpretable visualization tool to quantitatively assess trajectory plausibility and control consistency. Experiments demonstrate that our method achieves significantly lower MAE than Transformer, TCN, and conventional PID-plus-planning baselines on driving action prediction; it also exhibits superior generalization in interaction-intensive traffic scenarios. This work establishes a reproducible, data-driven paradigm for modeling real-world traffic interactions.
📝 Abstract
The Waymo Open Dataset has been released recently, providing a platform to crowdsource some fundamental challenges for automated vehicles (AVs), such as 3D detection and tracking. While the dataset provides a large amount of high-quality and multi-source driving information, people in academia are more interested in the underlying driving policy programmed in Waymo self-driving cars, which is inaccessible due to AV manufacturers’ proprietary protection. Accordingly, academic researchers have to make various assumptions to implement AV components in their models or simulations, which may not represent the realistic interactions in real-world traffic. Thus, this paper introduces an approach to learn a long short-term memory (LSTM)-based model for imitating the behavior of Waymo’s self-driving model. The proposed model has been evaluated based on Mean Absolute Error (MAE). The experimental results show that our model outperforms several baseline models in driving action prediction. In addition, a visualization tool is presented for verifying the performance of the model.
🔎 Similar Papers
No similar papers found.
