Lunar Robotic Operator for Autonomous Development of Surface Trails and Exploration Routes

Supervisor: Dr. William ‘Red’ Whittaker
Team: Ankit Aggarwal, Deepam Ameria, Bhaswanth Ayapilla, Simson D’Souza, Boxiang Fu
Project Website: Lunar ROADSTER
Project GitHub: Lunar ROADSTER GitHub

Humanity is preparing to return to the Moon, with the Artemis missions focusing on exploring the South Pole — a region rich in sites of interest. Establishing a circumnavigating route around the lunar pole will serve as a critical “highway” connecting these sites and enabling key activities such as transportation, human settlement, and resource extraction.
A solar-powered rover capable of sun-synchronous circumnavigation could achieve perpetual operation by avoiding lunar sunsets. At high latitudes, this is feasible at low speeds. However, these assumptions rely on the terrain being flat and traversable, free from major topographical challenges. A mission to manipulate the lunar regolith in the circumnavigating path to make it more traversable for future missions is thus, a clear step forward. A robotic system can be designed to conduct these operations efficiently for extended durations.
The Lunar Robotic Operator for Autonomous Development of Surface Trails and Exploration Routes (Lunar ROADSTER) is an autonomous moon-working rover, capable of finding exploration routes and grooming the lunar surface to develop traversable surface trails. These groomed trails will become the backbone for the colonization of the Moon by enabling transportation, logistics, and enterprise development.

My Contributions:
Built an Autonomous Moon-working Mechatronic Rover capable of grooming the lunar surface using a bulldozing mechanism
Developed rover autonomy using convex optimization (GLOP) on terrain point cloud maps for traversable path planning in ROS2
Fused ZED stereo and IMU data via an extended Kalman Filter for autonomous navigation (< 5 % pose error) and manipulation
Trained a YOLO v8 model to extract crater geometry using RGB-D data and validated using a point cloud-based gradient map

Project Video

Public Poster

Links

Detailed Report: Report
Full System Implementation: System Implementation
Full System Performance: System Performance
Media Gallery: Media