Learning to Predict Vehicle Trajectories with Model-based Planning

Song, Haoran; Luan, Di; Ding, Wenchao; Wang, Michael Yu; Chen, Qifeng

doi:10.48550/arxiv.2103.04027

Cited by 10 publications

(10 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another interesting observation is that methods performing very well on minFDE 6 such as LaneGCN [14] and TPCN [31] have a worse MR 6 as drawback. PRIME [28] has the closest MR 6 to ours but a much higher minFDE 6 in comparison. We show the results of both our sampling optimized for MR and minFDE with the same trained model.…”

Section: B Comparison With State-of-the-artmentioning

confidence: 42%

See 1 more Smart Citation

HOME: Heatmap Output for future Motion Estimation

Gilles¹,

Sabatini²,

Tsishkou³

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, we propose HOME, a framework tackling the motion forecasting problem with an image output representing the probability distribution of the agent's future location. This method allows for a simple architecture with classic convolution networks coupled with attention mechanism for agent interactions, and outputs an unconstrained 2D topview representation of the agent's possible future. Based on this output, we design two methods to sample a finite set of agent's future locations. These methods allow us to control the optimization trade-off between miss rate and final displacement error for multiple modalities without having to retrain any part of the model. We apply our method to the Argoverse Motion Forecasting Benchmark and achieve 1 st place on the online leaderboard.

show abstract

Section: B Comparison With State-of-the-artmentioning

confidence: 42%

“…Another family of methods use a pool of anchor trajectories, predefined [4] or model-based [23,28], and rank them with a learned model. This allows to avoid any mode collapse and assert realistic trajectories, but removes the ability to tune the trajectories accurately to the current situation.…”

Section: Related Workmentioning

confidence: 99%

HOME: Heatmap Output for future Motion Estimation

Gilles¹,

Sabatini²,

Tsishkou³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Anchor trajectories can also be regarded as a set of driving modes. MultiPath [6] clusters a set of anchors using offline data, while CoverNet [8] and PRIME [22] generate the candidate trajectory set according to the target's state and driving context in an online fashion. Goal-driven methods have also gained popularity in recent years.…”

Section: Related Workmentioning

confidence: 99%

Trajectory Prediction with Graph-based Dual-scale Context Fusion

Zhang

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

Motion prediction for traffic participants is essential for a safe and robust automated driving system, especially in cluttered urban environments. However, it is highly challenging due to the complex road topology as well as the uncertain intentions of the other agents. In this paper, we present a graph-based trajectory prediction network named the Dual Scale Predictor (DSP), which encodes both the static and dynamical driving context in a hierarchical manner. Different from methods based on a rasterized map or sparse lane graph, we consider the driving context as a graph with two layers, focusing on both geometrical and topological features. Graph neural networks (GNNs) are applied to extract features with different levels of granularity, and features are subsequently aggregated with attention-based inter-layer networks, realizing better local-global feature fusion. Following the recent goaldriven trajectory prediction pipeline, goal candidates with high likelihood for the target agent are extracted, and predicted trajectories are generated conditioned on these goals. Thanks to the proposed dual-scale context fusion network, our DSP is able to generate accurate and human-like multi-modal trajectories. We evaluate the proposed method on the large-scale Argoverse motion forecasting benchmark, and it achieves promising results, outperforming the recent state-of-the-art methods.

show abstract

“…However, they use the same extracted feature vector to output multiple trajectories, which is not intuitive and lacks interpretability on the outputs. Anchor-based methods [14], [15], [16] can provide better interpretability, feasibility, and diversity on the results, but their predictions are restricted to a predefined set, obtained by clustering from the data or generated by a model, which may impede the prediction accuracy. On the other hand, goal-based [4] or proposal-based methods [17], [18], [19] have been widely used due to their superior accuracy and interpretability.…”

Section: B Multi-modal Predictionmentioning

confidence: 99%

Multi-modal Motion Prediction with Transformer-based Neural Network for Autonomous Driving

Huang¹,

Mo²,

Lv³

2021

Preprint

View full text Add to dashboard Cite

Predicting the behaviors of other agents on the road is critical for autonomous driving to ensure safety and efficiency. However, the challenging part is how to represent the social interactions between agents and output different possible trajectories with interpretability. In this paper, we introduce a neural prediction framework based on the Transformer structure to model the relationship among the interacting agents and extract the attention of the target agent on the map waypoints. Specifically, we organize the interacting agents into a graph and utilize the multi-head attention Transformer encoder to extract the relations between them. To address the multi-modality of motion prediction, we propose a multimodal attention Transformer encoder, which modifies the multihead attention mechanism to multi-modal attention, and each predicted trajectory is conditioned on an independent attention mode. The proposed model is validated on the Argoverse motion forecasting dataset and shows state-of-the-art prediction accuracy while maintaining a small model size and a simple training process. We also demonstrate that the multi-modal attention module can automatically identify different modes of the target agent's attention on the map, which improves the interpretability of the model.

show abstract

Learning to Predict Vehicle Trajectories with Model-based Planning

Cited by 10 publications

References 45 publications

HOME: Heatmap Output for future Motion Estimation

HOME: Heatmap Output for future Motion Estimation

Trajectory Prediction with Graph-based Dual-scale Context Fusion

Multi-modal Motion Prediction with Transformer-based Neural Network for Autonomous Driving

Contact Info

Product

Resources

About