Publication Type : Conference Proceedings
Publisher : IEEE
Source : 2025 International Conference on Robotics and Mechatronics (ICRM)
Url : https://doi.org/10.1109/icrm66809.2025.11349097
Campus : Chennai
School : School of Engineering
Department : Mechanical Engineering
Year : 2025
Abstract : The goal of this research is to create precise path-following capabilities for the wind-powered seeding robot inspired by tumbleweed that is intended for field and greenland restoration in difficult conditions. Because of the chaotic physics of actual wind conditions, traditional methods frequently fail. Mirror Descent Guided Policy Search (MDGPS) is a model-based reinforcement learning algorithm that is used in the suggested solution. To efficiently control the robot, this algorithm trains a deep neural network using low-dimensional sensor data, like path deviation. By temporarily moving its internal mass to one side, the robot is able to steer. Compared to complicated aerodynamic modelling or conventional methods, this approach is noticeably more effective. Achieving accurate path-following even in extremely windy conditions is the main objective.The robot is designed to sow seeds in a unique manner by maneuvering to the desired direction. © 2025 IEEE.
Cite this Research Publication : Thanush J, K. Vyshnavi, Naren Narayana R, Naveen S K, Piyush Pratap Singh, Tumbleweed-Inspired Autonomous Seeding Robot with Tensegrity Structure and Reinforcement Learning Control, 2025 International Conference on Robotics and Mechatronics (ICRM), IEEE, 2025, https://doi.org/10.1109/icrm66809.2025.11349097