Experience

10X · IITM Research Park · May 2023 – Jul 2023

Autonomous Systems Engineer

An autonomous wheelchair built on IMU, encoders and 2D LiDAR with ROS, running Hector and Cartographer SLAM, alongside a multilingual RASA chatbot. Demonstrated to Google India's CEO and the Chief Justice of India. The visualisation below follows one full mapping run.

2D LiDAR SLAM pipeline from range scan to path plan

01 The floor plan

A map not yet mapped

The walls of the space, the loop the wheelchair will drive, and the robot at its start with the LiDAR wedge glowing. The occupancy grid is empty; the legend of free, occupied and unknown cells waits at the corner.

02 The mapping run

Drive, match, close the loop

Rays sweep and strike walls into a point cloud, ICP registers each scan, and the occupancy grid fills behind the robot. The pose graph accumulates drift until the loop-closure constraint snaps the trajectory straight, and an A* path is planned to the goal.

About this visualisation

The run drawn above: a 2D LiDAR ray-cast against the floor plan, ICP scan matching registering each sweep, and a log-odds occupancy grid of free, occupied and unknown cells filling as the robot drives. A Cartographer-style pose graph of submap nodes and odometry edges accumulates drift, a loop-closure constraint pulls the graph consistent, and A* plans a path through the corridor to the goal.

The first scroll holds the floor plan at rest. The second replays the run: everything the wheelchair's navigation stack did in real time, laid out step by step.