Autonomous Robot Navigation Using Learning Based Technology

Autonomous Mobile Robots & Vehicles

AI agents that learn to move naturally and respectfully around people by adapting in real timeensuring smooth, socially aware navigation in dynamic spaces

Vector (2)-da6325

Institute:
IIIT Allahabad

Vector (3)-42d8ac

PI Name:
Prof. G C Nandi

Technology Readiness Level (TRL)
4

Intellectual Property: PA No. 202311009245

Problem
Addressed

  • End to end navigation in dynamic and static environment using learning based technology.
  • Study of multiple behaviour of humans collision avoidance
  • Improved the directivity of the agent using learning based technology.
  • Self-imitating ability to human by robot.
Arrow 6

About the
Technology

This work develops socially aware navigation for AI agents using Expert Session-based Co-teaching to Reinforcement Learning (ESC-RL), blending human guidance with reinforcement learning for dynamic obstacle avoidance and social etiquette. Generative Adversarial Imitation Learning (GAIL) enhances adaptability by iteratively learning from human interventions, enabling efficient, real-world navigation without reliance on static datasets or rigid behavior models.
  • Supervised learning.
  • Reinforcement learning.
  • Generative Adversarial Imitation Learning
  • Heuristic property of path planning.

Application Areas & Use Cases

  • Autonomous navigation in structured(ware house) environment.
  • Social navigation in unstructure environment (Public places).
Arrow 6
Technology Name Autonomous mobile robot
SPurpose/Function Navigate dynamic environments safely and efficiently
Key Features Dynamic obstacle avoidance     • Reactive navigation     • Social behavio
Technical Details: Sensors: LiDAR . ROS-based
Target Environment Public spaces like airports, malls, hospitals and warehouse
Input Raw Lidar Data
Output Velocity