LaneRCNN: Distributed Representations for Graph-Centric Motion Forecasting

Zeng, Wei; Liang, Ming; Liao, Renjie; Urtasun, Raquel

doi:10.48550/arxiv.2101.06653

Cited by 10 publications

(33 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then a global interaction graph is applied to perform feature fusion among subgraphs. Meanwhile, LaneGCN [10] further utilizes the inherent topology of maps and proposes a novel lane convolution operator, achieving more effective context fusion, and LaneRCNN [14] 1 https://github.com/HKUST-Aerial-Robotics/DSP presents a graph-based representation for agents and conducts feature fusion in a graph-to-graph manner.…”

Section: Related Workmentioning

confidence: 99%

“…Goal-driven methods have also gained popularity in recent years. TNT [12] and LaneRCNN [14] perform endpoint classification and offset regression to generate multi-modal goals, followed by a completion network to get full trajectories. HOME [13] leverages CNNs to encode rasterized BEV images and output a heatmap which represents the probability distribution of the target's goals.…”

Section: Related Workmentioning

confidence: 99%

“…For example, trajectories and lanes are simply represented using polylines and further encoded into sparse nodes, making it hard to capture the local information. In the meantime, goal-driven methods [11]- [14] have achieved higher performance on various benchmarks. These methods roughly decompose the trajectory forecasting problem into two sub-tasks, namely, forecasting possible endpoints/goals of target agents, and completing full trajectories conditioned on both context features and predicted goals.…”

Section: Introductionmentioning

confidence: 99%

“…Currently, candidate goals are either represented as pixels on rasterized BEV images [13,15] or generated w.r.t. sparse lane graphs [12,14], making it difficult to consider both high-level and fine-grained driving context. Thus, it remains an interesting and challenging problem on how to construct a network that is capable of capturing multi-scale geometrical and topological features while being compatible with the goal-driven pipeline in a unified way.…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, the proposed method naturally adapts to the goal-driven prediction framework since the sampled nodes in the DA layer can be directly used as the potential goal candidate. Compared to the existing state-of-the-arts based on rasterized BEV images [5,13] and sparse lane graphs [10,12,14], our DSP is able to consider both lane topology and the fine-grained local features, resulting in better prediction performance.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

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

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Trajectory Prediction with Graph-based Dual-scale Context Fusion

Zhang

Chen³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

Deep learning-based interaction-aware trajectory prediction for autonomous vehicles

Mo¹

View full text Add to dashboard Cite

xv strong scene adaptability. Besides, the algorithms developed based on the proposed HGS pooling technique and the HEAT network won the championships of two worldwide autonomous vehicle prediction challenges, respectively. These outcomes demonstrate the feasibility and e↵ectiveness of the proposed methods. In addition, the high-level algorithm architectures, methodologies employed, and models developed in this work will expand the current theories of autonomous driving and intelligent transportation systems. They can also be expanded to a wide range of robotics and automation applications.

show abstract

On complementing end-to-end human behavior predictors with planning

Sun¹,

Jia²,

Dragan³

2021

Preprint

View full text Add to dashboard Cite

High capacity end-to-end approaches for human motion prediction have the ability to represent subtle nuances in human behavior, but struggle with robustness to out of distribution inputs and tail events. Planning-based prediction, on the other hand, can reliably output decentbut-not-great predictions: it is much more stable in the face of distribution shift, but it has high inductive bias, missing important aspects that drive human decisions, and ignoring cognitive biases that make human behavior suboptimal.In this work, we analyze one family of approaches that strive to get the best of both worlds: use the end-to-end predictor on common cases, but do not rely on it for tail events / out-of-distribution inputs -switch to the planningbased predictor there. We contribute an analysis of different approaches for detecting when to make this switch, using an autonomous driving domain. We find that promising approaches based on ensembling or generative modeling of the training distribution might not be reliable, but that there very simple methods which can perform surprisingly wellincluding training a classifier to pick up on tell-tale issues in predicted trajectories.

show abstract

LaneRCNN: Distributed Representations for Graph-Centric Motion Forecasting

Cited by 10 publications

References 0 publications

Trajectory Prediction with Graph-based Dual-scale Context Fusion

Trajectory Prediction with Graph-based Dual-scale Context Fusion

Deep learning-based interaction-aware trajectory prediction for autonomous vehicles

On complementing end-to-end human behavior predictors with planning

Contact Info

Product

Resources

About