Lunar Exploration Through Cooperative Rovers using Smart Navigation

Summary

 

This project introduces a multi-rover system equipped with advanced Guidance, Navigation, and Control (GNC) and a communication architecture designed for lunar exploration. The multi-rover system incorporates a centralized base station coordinating rovers to maximize lunar exploration within the network constraint. A key feature of this system is the integration of autonomous capabilities in the rovers, enhanced by visual SLAM for rover localization. The model employs Model Predictive Control (MPC) to navigate the rovers along predefined paths, incorporating a range of constraints including limitations on wheel angular velocities, communication range restrictions, and the necessity for crater avoidance. To test and validate the proposed system, simulations are conducted within the Gazebo simulation environment, fully integrated with the ROS framework, effectively replicating the physical behavior of the rovers. The simulation encompasses three distinct case studies. The study involves navigating a rover from the base station to the periphery of the safe communication range. Subsequently, the rovers are guided along circular trajectories, designed to ensure continuous communication both among the rovers and between the rovers and the base station. Through these simulations, the project demonstrates the efficacy of the proposed framework in a lunar exploration context.

​

This project proposes a novel approach for lunar surface exploration, leveraging a base control station on the lunar surface with a cooperative multi-rover system to maximize the exploration within limited range constraints. Wireless communication is an essential part of the data transfer and reception between the rovers and the base station. Instead of traditional “phoning home” methods of communication, the communication between the base station (lander) and rovers takes place within the lunar surface, saving significant time which facilitates the expedition to much farther planets where phoning home is not an optimal solution. Our system comprises two identical rovers, each tasked to maximize the area of the lunar terrain within the constraints of available network range, and a base control station acting as the central hub of the mission. A circular region around the station is conceptualized as the operational area with its radius defined by the station’s maximum network range. Within this area, a smaller circle with radius 𝑟1, a conservative estimate below the maximum range, marks the primary exploration zone for Rover 1. Both rovers maintain direct communication with the station inside this designated region. Throughout its journey, the rovers are tasked with creating a detailed visual and 3D map of the surroundings and simultaneously estimating its absolute position with respect to the station, employing Real Time-Based Mapping (RTAB-Map) using a stereo camera. Navigation is guided by the Rapidly-exploring Random Tree Star (RRT*) algorithm, enabling the rovers to efficiently maneuver towards the periphery of a circular region while avoiding lunar craters and hazards on the way. The rovers employ MPC for navigation, maneuvering around the station along a predetermined path and avoiding craters. Simultaneously, the secondary rover (Rover 2) extends its exploration to the edge of the larger circle, defined by the maximum range of the station’s network relayed by Rover 1. Rover 2 consequently circumnavigates the larger circle with a radius, 𝑟2, ensuring larger area exploration, while remaining in a distance constraint with Rover 1 to avoid loss of network. The performance of the proposed navigation strategy will be evaluated through simulation in Gazebo simulated lunar environment to determine its effectiveness in maximizing the exploration of the lunar surface.

​

​

​​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​