Learning Lane Graph Representations for Motion Forecasting

Liang, Ming; Yang, Bin; Hu, Rui; Chen, Yun; Liao, Renjie; Feng, Shijie; Urtasun, Raquel

doi:10.48550/arxiv.2007.13732

Cited by 16 publications

(38 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [25], instead of the Cartesian coordinate, the Frenét coordinate frame is employed to represent vehicles' states. Recently, [18], [33] proposed the use of Graph Neural Networks and for scene semantics to be represented as connected graphs. Their methods require the input to be represented in a vectorized format which is then converted to their own defined graph representation.…”

Section: Previous Workmentioning

confidence: 99%

“…Leveraging graph neural networks, they learn the relations between nodes and the final prediction. "LaneGCN" [33] constructs a lane graph from vectorized scene and learns interactions between the lanes and the agents.…”

Section: B Dataset and Metricsmentioning

confidence: 99%

“…Second, while the scene information suffices to understand the geometry of the scene, images are of high dimensions and potentially inefficient to learn from. To mitigate these issues, researchers recently proposed direct use of vectorized data as inputs [18], [33]. However, due to the use of a non-standard data format, these approaches are limited to their own model designs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

SVG-Net: An SVG-based Trajectory Prediction Model

Bahari,

Zehtab,

Khorasani

et al. 2021

Preprint

View full text Add to dashboard Cite

Anticipating motions of vehicles in a scene is an essential problem for safe autonomous driving systems. To this end, the comprehension of the scene's infrastructure is often the main clue for predicting future trajectories. Most of the proposed approaches represent the scene with a rasterized format and some of the more recent approaches leverage custom vectorized formats. In contrast, we propose representing the scene's information by employing Scalable Vector Graphics (SVG). SVG is a well-established format that matches the problem of trajectory prediction better than rasterized formats while being more general than arbitrary vectorized formats. SVG has the potential to provide the convenience and generality of raster-based solutions if coupled with a powerful tool such as CNNs, for which we introduce SVG-Net. SVG-Net is a Transformer-based Neural Network that can effectively capture the scene's information from SVG inputs. Thanks to the selfattention mechanism in its Transformers, SVG-Net can also adequately apprehend relations amongst the scene and the agents. We demonstrate SVG-Net's effectiveness by evaluating its performance on the publicly-available Argoverse forecasting dataset. Finally, we illustrate how, by using SVG, one can benefit from datasets and advancements in other research fronts that also utilize the same input format. Our code is available at https://vita-epfl.github.io/SVGNet/.

show abstract

Section: Previous Workmentioning

confidence: 99%

Section: B Dataset and Metricsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SVG-Net: An SVG-based Trajectory Prediction Model

Bahari,

Zehtab,

Khorasani

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Several works [3,18,22,25] attempted to extract and reconstruct road topologies from overhead images. VectorNet [14] and LaneGCN [23] proposed to efficiently encode roads using graph attention networks and graph convolutions, replacing traditional rendering-based models. Chu et al propose Neural Turtle Graphics (NTG) [10], a generative model to produce roads iteratively.…”

Section: Street Map Modeling and Generationmentioning

confidence: 99%

HDMapGen: A Hierarchical Graph Generative Model of High Definition Maps

Lu¹,

Zhao²,

Nash³

et al. 2021

Preprint

View full text Add to dashboard Cite

High Definition (HD) maps are maps with precise definitions of road lanes with rich semantics of the traffic rules. They are critical for several key stages in an autonomous driving system, including motion forecasting and planning. However, there are only a small amount of realworld road topologies and geometries, which significantly limits our ability to test out the self-driving stack to generalize onto new unseen scenarios. To address this issue, we introduce a new challenging task to generate HD maps. In this work, we explore several autoregressive models using different data representations, including sequence, plain graph, and hierarchical graph. We propose HDMapGen, a hierarchical graph generation model capable of producing high-quality and diverse HD maps through a coarse-to-fine approach. Experiments on the Argoverse dataset and an inhouse dataset show that HDMapGen significantly outperforms baseline methods. Additionally, we demonstrate that HDMapGen achieves high scalability and efficiency.

show abstract

“…The disadvantages also include encoding a quite large amount of irrelevant to the underlying process information from the entire scene. To overcome these drawbacks most state-of-the-art approaches lean on the graph structure [3], [4]. Graph-based methods take into account only relevant to the driving patterns information and can provide better quality due to the expressive power of graph neural networks [5], [6].…”

Section: Introductionmentioning

confidence: 99%

Estimating Uncertainty For Vehicle Motion Prediction on Yandex Shifts Dataset

Pustynnikov¹,

Eremeev²

2021

Preprint

View full text Add to dashboard Cite

Motion prediction of surrounding agents is an important task in context of autonomous driving since it is closely related to driver's safety. Vehicle Motion Prediction (VMP) track of Shifts Challenge ** focuses on developing models which are robust to distributional shift and able to measure uncertainty of their predictions. In this work we present the approach that significantly improved provided benchmark and took 2nd place on the leaderboard.

show abstract

Learning Lane Graph Representations for Motion Forecasting

Cited by 16 publications

References 40 publications

SVG-Net: An SVG-based Trajectory Prediction Model

SVG-Net: An SVG-based Trajectory Prediction Model

HDMapGen: A Hierarchical Graph Generative Model of High Definition Maps

Estimating Uncertainty For Vehicle Motion Prediction on Yandex Shifts Dataset

Contact Info

Product

Resources

About