Less is More: Non-uniform Road Segments are Efficient for Bus Arrival Prediction

TL;DR

This work tackles bus arrival prediction by moving away from uniform road segmentation and learning informative, non-uniform segments through a Deep Progressive Reinforcement Learning framework. The method decouples prediction into a lightweight linear ETA model and a Road Segment Selection Network (RSS-Net) that adaptively selects segments via policy-gradient optimization with a reward balancing accuracy and sparsity. Across large-scale real-world data, the approach matches or surpasses state-of-the-art nonlinear baselines while using far fewer input features, improving computational efficiency and real-time applicability. The results demonstrate robust generalization across routes and functional zones, highlighting the practical impact of non-uniform segmentation for traffic prediction tasks.

Abstract

In bus arrival time prediction, the process of organizing road infrastructure network data into homogeneous entities is known as segmentation. Segmenting a road network is widely recognized as the first and most critical step in developing an arrival time prediction system, particularly for auto-regressive-based approaches. Traditional methods typically employ a uniform segmentation strategy, which fails to account for varying physical constraints along roads, such as road conditions, intersections, and points of interest, thereby limiting prediction efficiency. In this paper, we propose a Reinforcement Learning (RL)-based approach to efficiently and adaptively learn non-uniform road segments for arrival time prediction. Our method decouples the prediction process into two stages: 1) Non-uniform road segments are extracted based on their impact scores using the proposed RL framework; and 2) A linear prediction model is applied to the selected segments to make predictions. This method ensures optimal segment selection while maintaining computational efficiency, offering a significant improvement over traditional uniform approaches. Furthermore, our experimental results suggest that the linear approach can even achieve better performance than more complex methods. Extensive experiments demonstrate the superiority of the proposed method, which not only enhances efficiency but also improves learning performance on large-scale benchmarks. The dataset and the code are publicly accessible at: https://github.com/pangjunbiao/Less-is-More.

Figures (6)

• Figure 1: Road segments significantly affect the bus ETA prediction. These spots have different temporal patterns to cause diverse predictability of bus ETA prediction.
• Figure 2: Prediction error is propagated into next prediction in auto-regression method.
• Figure 3: The architecture of the proposed method. It depicts the operational pipeline of the RSS-Net and LRM during the training and testing phases.
• Figure 4: The neural network architecture of RSS-Net for adaptive segment selection.
• Figure 5: Comparison of MAE performance under different proportions of selected interpolation points.

Theorems & Definitions (3)

• Definition 1: A journey of a bus
• Definition 2: Road segment
• Definition 3: The ETA prediction on sparse road segments