research

Publications and research positions — cooperative mobile manipulation, Ackermann autonomy at e-Yantra, SLAM and soft robotics.

research areaseach links to the matching projects

Navigation

Ackermann, differential and holonomic platforms. Nav2, global and local planners, pure pursuit.

SLAM & Estimation

Visual SLAM written from scratch; 3D LiDAR SLAM flown on a drone. EKF, AMCL, pose-graph optimization.

Perception

Depth cameras, ArUco pose, homography, bird's-eye-view lane pipelines.

Manipulation

IK, Jacobians and null-space control, visual servoing, QP whole-body control.

Learning

ACT, diffusion policy, SmolVLA, π0 — action chunking and flow matching for manipulation.

Hardware

ESP32 low-level control, power distribution, Jetson, LiDAR and depth-camera integration.

publications

2026

Published

Soft autonomous mobile manipulators in agricultural automation – a review

Tahsin Khan, Mervin Joe Thomas

Industrial Robot, Vol. 53 No. 2, pp. 376–398

Soft grippers, actuation and control, localization and motion planning for agricultural manipulators — and why integrating them, rather than any one of them, is the open problem.

  • Soft Robotics
  • Mobile Manipulation
  • Review

research experience

Cooperative Mobile Manipulation of a Shared Payload

Researcher · IIT Bombay

  • 2–3 mobile manipulators carrying one payload, with no central controller and no external tracking.
  • Decentralized localization and on-board compute; local sensing only — base depth cameras and wrist cameras.
  • QP-based whole-body control across base and arm under the payload's closed-chain constraint. Learning-based coordination next.
  • Whole-Body Control
  • QP Control
  • Decentralized Estimation
  • Multi-Robot

e-Yantra Robotics Competition — Theme Development

Senior Project Technical Assistant · ERTS Lab, IIT Bombay

  • Build the competition themes end to end — reference implementation, simulation, hardware bring-up and the task progression students work through.
  • Niti Vahan: an Ackermann vehicle driving a scaled city from one forward camera. Built the perception pipeline — bird's-eye-view homography, colour segmentation, polynomial lane fitting, obstacle layer for lane switching.
  • Localization in progress: visual odometry with landmark correction, then autonomous parking. Previous theme: Krishi CoBot, an autonomous greenhouse arm-plus-base system.
  • Ackermann Autonomy
  • Visual Odometry
  • Lane Detection
  • ROS 2

Field Exploration with ROS-Powered Vehicles

Robotics Research Intern · Autonomous Robots and Multi-Systems (ARMS) Lab, IIT Bombay

Advised by Prof. Arpita Sinha

  • Built a ROS 2-controlled 4WD Ackermann vehicle for field exploration — converted from an RC car, working mostly on the hardware side: electronics, circuit design and ESP32 low-level steering and drive control.
  • Implemented a Kalman filter fusing IMU data with Vicon velocity estimates for accurate odometry.
  • Used a pure-pursuit motion planner for autonomous waypoint navigation.
  • Performed LiDAR SLAM and leveraged Nav2 AMCL for localization, with a custom motion planner for navigation.
  • ROS 2
  • Kalman Filter
  • LiDAR SLAM
  • Nav2 AMCL
  • Pure Pursuit
  • ESP32

Soft Autonomous Mobile Manipulators in Agriculture

Undergraduate Researcher · NITK Surathkal

Advised by Prof. Mervin Joe Thomas

  • Reviewed soft gripping and actuation, mobile-base autonomy and motion planning for agricultural harvesting.
  • Published in Industrial Robot (2026).
  • Soft Grippers
  • Mobile Manipulators
  • Review