|
Sanjeev Kannan
I am currently pursuing a Masters in Robotics Engineering at Worcester Polytechnic Institute (WPI). I am an aspiring Roboticist with interests in Motion Planning and Control. I have experience working with autonomous mobile robots and multi-robot systems.
|
Email: skannan[at]wpi.edu
Resume /
GitHub / LinkedIn / Hackerrank
|
Relevant Work
Motion Planning and Control of an Autonomous Car using Hybrid A*
|
|
Kinematic contraints of a 4-wheel vehicle are taken into consideration to plan and execute paths. Hybrid A* is used for high-level path planning (shown in Fig) and PID control is used for lower-level control. The project has been done using ROS-Gazebo, the Husky 4-wheeled UGV package and programmed in C++.
|
Quadrotor Motion Planning
|
|
In this work , we create and integrate planning, control and obstacle avoidance modules for a quadrotor. The quadrotor is successfully able to pass through obstacle rings, achieve a desired route and return to origin. We used the RRT-star algorithm to create the planning module.
|
Multi-Robot Graph Coverage
|
|
Here, an algorithm for collaborative exploration of a graph environment is implemented. A group of decentralized robots autonomously explore the graph environment. Communication between robots is done using beacons dropped by robots at graph nodes. Mathematical formulation and proof for quickest time to explore graph, number of robots required and ideal robot starting positions are found.
|
PID Control for a Mobile robot
|
|
A PID controller is created and implemented for navigating in a Turtlebot 2 robot. Precise navigation is achieved both on ROS-Gazebo simulation and physical hardware implementation.
|
|
Two methods for 3D Point Cloud Registration are implemented and their performace is compared - ICP ( Iterative Closest Point ) and Go-ICP (Global optima ICP).
|
|