IARN

Indoor Autonomous Robot Navigation

Hospitals and Hospice Care

Parameters

• 6 Months
• Starts in August 2020
• ~15 collaborators

“Learning Navigation Behaviors End-to-End with AutoRL,” “PRM-RL: Long-Range Robotic Navigation Tasks by Combining Reinforcement Learning and Sampling-based Planning”, and “Long-Range Indoor Navigation with PRM-RL”

Primary Functions

• Point-to-Point Long-range Navigation
• Path Following / Guided Navigation
• Collision avoidance

Technologies

ROS

• Framework for simulating complex robot behavior
• Pose Estimation, Localization, and Navigation
• Transfer learned parameters to real world

SLAM

• Simultaneous localization and mapping
• Creating a map of environment and tracking robot position 
• ROS includes tools for SLAM - gmapping

SLAM

Robot Platform

TurtleBot 3 Burger

LiDAR Camera

Reinforcement Learning

Reinforcement Learning

• State : noisy LiDAR observations
• Action : Differential Drive Wheel Speeds
• Reward : Reach goal while avoiding obstacles

Secondary Tasks

Monitoring and Debugging

• Monitor and Log state-action information
• Real-time visualization of trajectories

Web/Mobile Interface

• Track the status of the robot
• Sensors, Battery, etc,
• Control 

On-board UI

• Secure Access
• Control
• Networking
• Intervention/Reset