Off-Road Autonomous Navigation
Team lead for India's first off-road autonomous navigation system on a tracked platform (DRDO).
Problem
Tracked platforms operating in off-road conditions — jungles, rocky terrain, slopes, trenches, and adverse weather (heavy rain, fog, dust storms) — require autonomy stacks far beyond structured indoor AMR navigation.
Approach
As team lead, I led development of the first off-road autonomous navigation system (ANS) for a tracked platform with DRDO:
- Multi-sensor fusion — LiDAR, EO/IR camera, IMU
- Deep learning object detection for obstacle recognition
- 2.5D elevation mapping for traversability and negative obstacle avoidance
- Model predictive control for motion on uneven terrain
- ROS 2 software framework for modular integration and testing
Results
- Project completed by August 2025
- First off-road ANS of its kind for a tracked platform in India
- Validated across jungle, rocky, slope, trench, and desert conditions with adverse weather
Technologies
ROS 2 · LiDAR · EO/IR · IMU · Deep learning · Elevation mapping · MPC · Sim2Real