Trajectron++: Dynamically-Feasible Trajectory Forecasting With Heterogeneous Data

Salzmann, Tim; Ivanovic, Boris; Chakravarty, Punarjay; Pavone, Marco

doi:10.48550/arxiv.2001.03093

Cited by 40 publications

(142 citation statements)

References 0 publications

Supporting

Mentioning

141

Contrasting

Order By: Relevance

“…3) Trajectron++ [4]: One of the latest works in this direction is named Trajectron++ [4] has four configurations of predictions, of which we used Most Likely, which gives the best results of ADE and FDE. Trajectron++ calculates the result of predictions as a Gaussian Mixture Model (GMM) witch contains 25 Gaussian distributions.…”

Section: Discussionmentioning

confidence: 99%

“…Broadly, human motion prediction can be divided in two classes: model-based that creates an empirical model of these transition functions such as the SFM [6] and its variants [18], [19] or models proposed from the graphics community [20], [21]. And Learning-based approaches [4], [7], [22]- [25] that are becoming the dominant paradigm in human motion prediction, as well as in other topics due to its unrivaled results. Both of these approaches have the same input and output data and predict the next state of the pedestrians in dt time.…”

Section: Related Workmentioning

confidence: 99%

“…Trajectron++ [4], [28] -a multi-agent behavior prediction model that accounts for the dynamics of the agents, produces predictions possibly conditioned on potential future robot trajectories which can effectively use heterogeneous data about the surrounding environment.…”

Section: Related Workmentioning

confidence: 99%

“…CovarianceNet in cyan stars, blue and green ellipses, -mean, 1 sigma, 3 sigma, respectively. Trajectron++ [4] in yellow stars, yellow and red ellipses, red ellipsesmean, 1 sigma, 3 sigma, respectively. The environment is challenging, since there is a pedestrian in the way.…”

Section: Introductionmentioning

confidence: 99%

“…2, consisting of deep model for Goals Prediction and model-based trajectory prediction, complemented by a modern neural-net approach to predict motion as a unimodal Gaussian distribution. Several works provide a multimodal distribution for motion prediction such as [4], achieving excellent results in benchmarks. However, combining a mixture distribution into robot planning approaches (for instance) requires careful considerations.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

CovarianceNet: Conditional Generative Model for Correct Covariance Prediction in Human Motion Prediction

Postnikov¹,

Gamayunov²,

Ferrer³

2021

Preprint

View full text Add to dashboard Cite

The correct characterization of uncertainty when predicting human motion is equally important as the accuracy of this prediction. We present a new method to correctly predict the uncertainty associated with the predicted distribution of future trajectories. Our approach, CovariaceNet, is based on a Conditional Generative Model with Gaussian latent variables in order to predict the parameters of a bi-variate Gaussian distribution. The combination of CovarianceNet with a motion prediction model results in a hybrid approach that outputs a uni-modal distribution. We will show how some state of the art methods in motion prediction become overconfident when predicting uncertainty, according to our proposed metric and validated in the ETH data-set [1]. CovarianceNet correctly predicts uncertainty, which makes our method suitable for applications that use predicted distributions, e.g., planning or decision making.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

CovarianceNet: Conditional Generative Model for Correct Covariance Prediction in Human Motion Prediction

Postnikov¹,

Gamayunov²,

Ferrer³

2021

Preprint

View full text Add to dashboard Cite

show abstract

Conditional Generative Adversarial Networks for Speed Control in Trajectory Simulation

Julka

Sowrirajan

Schlötterer

2022

Machine Learning, Optimization, and Data Science

View full text Add to dashboard Cite

Motion behaviour is driven by several factors -goals, presence and actions of neighbouring agents, social relations, physical and social norms, the environment with its variable characteristics, and further. Most factors are not directly observable and must be modelled from context. Trajectory prediction, is thus a hard problem, and has seen increasing attention from researchers in the recent years. Prediction of motion, in application, must be realistic, diverse and controllable. In spite of increasing focus on multimodal trajectory generation, most methods still lack means for explicitly controlling different modes of the data generation. Further, most endeavours invest heavily in designing special mechanisms to learn the interactions in latent space. We present Conditional Speed GAN (CSG), that allows controlled generation of diverse and socially acceptable trajectories, based on user controlled speed. During prediction, CSG forecasts future speed from latent space and conditions its generation based on it. CSG is comparable to state-ofthe-art GAN methods in terms of the benchmark distance metrics, while being simple and useful for simulation and data augmentation for different contexts such as fast or slow paced environments. Additionally, we compare the effect of different aggregation mechanisms and show that a naive approach of concatenation works comparable to its attention and pooling alternatives.

show abstract

Multi‐agent trajectory prediction with adaptive perception‐guided transformers

Nguyen,

Yoo

2024

IET Intelligent Trans Sys

View full text Add to dashboard Cite

The ability to predict the trajectory of an autonomous vehicle accurately is crucial for safe and efficient navigation. However, predicting diverse and multimodal futures can be challenging. Recent approaches such as attention and graph neural networks have achieved state‐of‐the‐art performance by considering agent interactions and map contexts. This study focused on multi‐agent prediction using an agent‐centric approach with transformers. This enables parallel computation and a comprehensive understanding of the environment. Two main features are introduced: an adaptive receptive field (ARF) that captures the relevant surroundings for each agent, and perception encoding, which serves as spatial context embeddings. The ARF adapts to the agent's velocity and rotation, focusing attention ahead at high speeds or to the sides when it is slower. Perception encoding divides agents or lanes into levels and encodes the information of each level. This approach enables the efficient encoding of complex spatial relationships. The proposed method combines these advances with transformer modelling for multi‐agent trajectory prediction while ensuring real‐time prediction capabilities. The approach is evaluated on the Argoverse benchmark and better performance than the state‐of‐the‐art baseline is achieved. By addressing challenges such as multimodal outputs and robustness, the study enhances the safety and efficiency of autonomous driving systems by more accurately predicting trajectories.

show abstract

Trajectron++: Dynamically-Feasible Trajectory Forecasting With Heterogeneous Data

Cited by 40 publications

References 0 publications

CovarianceNet: Conditional Generative Model for Correct Covariance Prediction in Human Motion Prediction

CovarianceNet: Conditional Generative Model for Correct Covariance Prediction in Human Motion Prediction

Conditional Generative Adversarial Networks for Speed Control in Trajectory Simulation

Multi‐agent trajectory prediction with adaptive perception‐guided transformers

Contact Info

Product

Resources

About