No RL, No Simulation: Learning to Navigate without Navigating

Keyphrases

• Reinforcement Learning 100%
• Simulation-based Learning 100%
• Navigation Task 40%
• Ground Truth Map 40%
• Learning-based 20%
• Highly Effective 20%
• Simulation Environment 20%
• Robotic Platform 20%
• Simple Questions 20%
• Real Domain 20%
• Online Policy 20%
• Supervised Approach 20%
• Active Interaction 20%
• Future Image 20%
• Visual Navigation 20%
• Domain Gap 20%
• Learning Navigation 20%
• Sim-to-real 20%
• Policy Interaction 20%

Computer Science

• Reinforcement Learning 100%
• Navigation Task 40%
• Simulation Environment 20%
• robotic platform 20%
• Interaction Policy 20%
• Navigation Policy 20%