Visual Navigation Among Humans With Optimal Control as a Supervisor

Published in

IEEE Robotics and Automation Letters(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

This work proposes an approach that combines learning-based perception with model-based optimal control to navigate among humans based only on monocular, first-person RGB images that can generalize to previously unseen environments and human behaviors, and transfer directly from simulation to reality.

Abstract

Real world visual navigation requires robots to operate in unfamiliar, human-occupied dynamic environments. Navigation around humans is especially difficult because it requires anticipating their future motion, which can be quite challenging. We propose an approach that combines learning-based perception with model-based optimal control to navigate among humans based only on monocular, first-person RGB images. Our approach is enabled by our novel data-generation tool, HumANav, that allows for photorealistic renderings of indoor environment scenes with humans in them, which are then used to train the perception module entirely in simulation. Through simulations and experiments on a mobile robot, we demonstrate that the learned navigation policies can anticipate and react to humans without explicitly predicting future human motion, generalize to previously unseen environments and human behaviors, and transfer directly from simulation to reality. Videos describing our approach and experiments, as well as a demo of HumANav are available on the project website.1

Authors

Varun Tolani

2 papers

Somil Bansal

1 papers

Aleksandra Faust

1 papers

References56 items

Deep Drone Acrobatics

Neural Topological SLAM for Visual Navigation

DenseCAvoid: Real-time Navigation in Dense Crowds using Anticipatory Behaviors

Learning by Cheating

A Hamilton-Jacobi Reachability-Based Framework for Predicting and Analyzing Human Motion for Safe Planning

Visual Navigation Among Humans With Optimal Control as a Supervisor

Published in

IEEE Robotics and Automation Letters(2020)

External Links:

Generate Graph DownloadPDF

TL;DR

Abstract

Authors

Varun Tolani

2 papers

Somil Bansal

1 papers

Aleksandra Faust

1 papers

References56 items

Deep Drone Acrobatics

Neural Topological SLAM for Visual Navigation

DenseCAvoid: Real-time Navigation in Dense Crowds using Anticipatory Behaviors

Learning by Cheating

A Hamilton-Jacobi Reachability-Based Framework for Predicting and Analyzing Human Motion for Safe Planning

JRDB: A Dataset and Benchmark for Visual Perception for Navigation in Human Environments

RL-RRT: Kinodynamic Motion Planning via Learning Reachability Estimators From RL Policies

End-To-End Interpretable Neural Motion Planner

Human motion trajectory prediction: a survey

Deep Local Trajectory Replanning and Control for Robot Navigation

Neural Autonomous Navigation with Riemannian Motion Policy

Habitat: A Platform for Embodied AI Research

Deep Visual MPC-Policy Learning for Navigation

Combining Optimal Control and Learning for Visual Navigation in Novel Environments

A Behavioral Approach to Visual Navigation with Graph Localization Networks

DIViS: Domain Invariant Visual Servoing for Collision-Free Goal Reaching

ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst

Getting Robots Unfrozen and Unlost in Dense Pedestrian Crowds

Learning Navigation Behaviors End-to-End With AutoRL

Crowd-Robot Interaction: Crowd-Aware Robot Navigation With Attention-Based Deep Reinforcement Learning

Computer vision and deep learning techniques for pedestrian detection and tracking: A survey

QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation

A Tour of Reinforcement Learning: The View from Continuous Control

Deep Drone Racing: Learning Agile Flight in Dynamic Environments

Gibson Env: Real-World Perception for Embodied Agents

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions

Motion Planning Among Dynamic, Decision-Making Agents with Deep Reinforcement Learning

Driving Policy Transfer via Modularity and Abstraction

Perception, Guidance, and Navigation for Indoor Autonomous Drone Racing Using Deep Learning

Semi-parametric Topological Memory for Navigation

DroNet: Learning to Fly by Driving

Aggressive Deep Driving: Combining Convolutional Neural Networks and Model Predictive Control

Intention-Net: Integrating Planning and Deep Learning for Goal-Directed Autonomous Navigation

PRM-RL: Long-range Robotic Navigation Tasks by Combining Reinforcement Learning and Sampling-Based Planning

Agile Off-Road Autonomous Driving Using End-to-End Deep Imitation Learning

Memory Augmented Control Networks

Role Playing Learning for Socially Concomitant Mobile Robot Navigation

Learning to fly by crashing

Virtual-to-real deep reinforcement learning: Continuous control of mobile robots for mapless navigation

Cognitive Mapping and Planning for Visual Navigation

Uncertainty-Aware Reinforcement Learning for Collision Avoidance

Learning from Synthetic Humans

(CAD)$^2$RL: Real Single-Image Flight without a Single Real Image

From perception to decision: A data-driven approach to end-to-end motion planning for autonomous ground robots

Target-driven visual navigation in indoor scenes using deep reinforcement learning

3D Semantic Parsing of Large-Scale Indoor Spaces

Learning Deep Object Detectors from 3D Models

The dynamic window approach to collision avoidance