It Is Not the Journey but the Destination: Endpoint Conditioned Trajectory Prediction (2020-04-04T00:00:00.000000Z)

TL;DR

This work presents Predicted Endpoint Conditioned Network (PECNet) for flexible human trajectory prediction, which improves state-of-the-art performance on the Stanford Drone trajectory prediction benchmark and on the ETH/UCY benchmark by ~40.8%.

Abstract

Human trajectory forecasting with multiple socially interacting agents is of critical importance for autonomous navigation in human environments, e.g., for self-driving cars and social robots. In this work, we present Predicted Endpoint Conditioned Network (PECNet) for flexible human trajectory prediction. PECNet infers distant trajectory endpoints to assist in long-range multi-modal trajectory prediction. A novel non-local social pooling layer enables PECNet to infer diverse yet socially compliant trajectories. Additionally, we present a simple "truncation-trick" for improving few-shot multi-modal trajectory prediction performance. We show that PECNet improves state-of-the-art performance on the Stanford Drone trajectory prediction benchmark by ~20.9% and on the ETH/UCY benchmark by ~40.8%. Project homepage: this https URL

Authors

Adrien Gaidon

7 papers

J. Malik

14 papers

K. Mangalam

6 papers

TL;DR

Abstract

Authors

References47 items

SimAug: Learning Robust Representations from 3D Simulation for Pedestrian Trajectory Prediction in Unseen Cameras

SimAug: Learning Robust Representations from Simulation for Trajectory Prediction

Trajectory Forecasts in Unknown Environments Conditioned on Grid-Based Plans

Conditional Flow Variational Autoencoders for Structured Sequence Prediction

Human motion trajectory prediction: a survey

Conditional Generative Neural System for Probabilistic Trajectory Prediction

PRECOG: PREdiction Conditioned on Goals in Visual Multi-Agent Settings

Large Scale GAN Training for High Fidelity Natural Image Synthesis

SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints

Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks

Understanding Human Behaviors in Crowds by Imitating the Decision-Making Process

Non-local Neural Networks

Pedestrian Prediction by Planning Using Deep Neural Networks

Probabilistic vehicle trajectory prediction over occupancy grid map via recurrent neural network

DESIRE: Distant Future Prediction in Dynamic Scenes with Interacting Agents

Enabling Flow Awareness for Mobile Robots in Partially Observable Environments

Learning Social Etiquette: Human Trajectory Understanding In Crowded Scenes

Social LSTM: Human Trajectory Prediction in Crowded Spaces

Generative Adversarial Imitation Learning

Knowledge Transfer for Scene-Specific Motion Prediction

Goal-Directed Pedestrian Prediction

Unsupervised robot learning to predict person motion

Pedestrian's Trajectory Forecast in Public Traffic with Artificial Neural Networks

Socially-Aware Large-Scale Crowd Forecasting

Generative Adversarial Nets

Will the Pedestrian Cross? A Study on Pedestrian Path Prediction

Who are you with and where are you going?

Mobile Agent Trajectory Prediction using Bayesian Nonparametric Reachability Trees

Improving Data Association by Joint Modeling of Pedestrian Trajectories and Groupings

Recognition of situation classes at road intersections

Action understanding as inverse planning

Planning-based prediction for pedestrians

Abnormal crowd behavior detection using social force model

Crowds by Example

Probabilistic Robotics (Intelligent Robotics and Autonomous Agents)

Learning Motion Patterns of People for Compliant Robot Motion

Voronoi tracking: location estimation using sparse and noisy sensor data

Learning motion patterns of persons for mobile service robots

Camera-based observation of obstacle motions to derive statistical data for mobile robot motion planning

Long Short-Term Memory

Social force model for pedestrian dynamics.

Trajnet: Towards a benchmark for human trajectory prediction

predic-PECNet: Pedestrian Endpoint Conditioned Trajectory Prediction Network 17 tion in crowded spaces

Activity forecasting

Research Showcase @ CMU

Modelling Smooth Paths Using Gaussian Processes

Kitani , Dariu M . Gavrila , and Kai O . Arras

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names