Select Page

Lunabotics: Team Selene

Overview

NASA encourages universities around the nation to compete in a robotic mining competition to inspire innovative concepts and designs for the space resource industry. Proposed solutions centered on in-situ resource utilization are economically feasible as lunar exploration and development continues. 

In order to have a sustainable presence on the Moon, it is necessary to excavate lunar regolith to extract local resources. Lunabotics is a college level event that supports the Artemis Moon to Mars mission requiring teams to participate in multiple events throughout the year. The goal of this Artemis challenge is to gather and evaluate excavator design and surface motion operational robotic data for future lunar excavators.

Live Zoom Chat

Click the button below to join us live from 8:00 – 10:30 a.m. (MDT) on April 29.

Team Members

  • Nadia Scharunovych
  • Zeyun Yang
  • Aneesh Pawar
  • Daniel Abel
  • Joe Haynes
  • Jason Mackay
  • Ryan Weber
  • Marcus Brown

The Client

  • Angel Abbud-Madrid
  • NASA

Acknowledgements

Project Advisor: Marcus Florida

Technical Advisor: Prof.  Xiaoli Zhang

Video

Elevator Pitch

Colorado School of Mines RMC Team, Team Selene, is a part of NASA’s 2021 Robotic Mining Challenge aiming to construct a robot that can navigate from a starting point, traverse through the obstacle field, arrive at a mining area, excavate subsurface material, and return that material to the starting point for delivery to a collective sieve. Team Selene’s robot is composed of five subsystems, Drivetrain/Chassis, Excavation, Collection/Dumping, Control/Navigation, and Power. The robot is unique in the excavation wheel being centered in the chassis and parallel to the length of the robot. Our aim is to create a functional robot to compete in the 2021 competition.

Design Approach

The NASA-provided RMC rules and scoring rubric became the primary driver for subsystem prioritization and performance optimization. After considering how many points could be earned and lost in each scoring category, the team subsequently prioritized the excavation system, average data transmission bandwidth, and movement aspects of the robot. Weight minimization and energy consumption became secondary points of concern. Tertiary or minimal concerns were autonomy and dust mitigation which factored the least heavily into the team’s prioritization scheme. Automation was left as the final point of focus, only to be worked on if the other aspects of the robot are performing at the desired level. This was due to many pandemic-related setbacks stalling the completion of the project.

 

 

Design Solution

Mechanical Subsystem

Excavation:

The arm works as a first class lever, powered by actuators, to lower and raise the excavation subsystem during the mining portion of the competition. This design was chosen due to the increased degree of movement that the arm can lower to reach the goal depth in the regolith. The excavation wheel is powered by a single motor and includes a wheel insert that is separate from the wheels rotation to allow for the moon regolith to travel to the conveyor belt. The orientation of the wheel allows for a more space efficient design and allows for the wheel to be mounted securely to the excavation arm.

Chassis:

The drivetrain is composed of the wheels and motors. The drivetrain implementation is four-wheel drive with fixed wheels and independent axles. The design has been proven in previous competitions and is simple to implement. The independence of the wheels provides skid steering an added benefit that will be tested with respect to autonomy. The chassis is built to support the excavation subsystem and store the hardware. The chassis has been redesigned to optimize the bounding box space more efficiently and has been welded to offer a more concrete and sturdier skeleton to build off. The chassis now follows the perimeter of the bounding box and offers the main subsystems protection from any possible collisions while traversing the lunar terrain.

Conveyor Belt & Hopper:

The transportation of moon regolith to the conveyor belt is fixed by a chute that can be implemented into the wheel and attached to the conveyor belt. The second and third aspects of the excavation subsystem is the conveyor belt and the bucket. The wheel and bucket are inseparable in design. The current implementation has the belt move the excavated load to the bucket which is then moved using actuators to dump in the competition hopper. The excavation subsystem features a longer conveyor belt that dumps into a mounted bucket, the regolith is then released onto the dumping area with a linear actuator that opens a back door to the bucket. This design counteracts the force of the front excavation wheel and centers the weight in the back. In addition, this design makes autonomy easier to implement and ensures that the dumped regolith is accurate and precise. The conveyor belt offers a unique design, which follows the movement of the wheel to collect the excavated moon regolith. The smaller conveyor belt allows for flexibility in its positioning and dumps directly into the hopper for further exploration.

Electrical Subsystem

Wireless Communication:

The wireless communication system begins with the user side. On the user side, a Playstation 3 controller is wirelessly connected to the offboard ESP32 through bluetooth. The offboard ESP32 collects sensor data for two axes of each joystick, two trigger values, and four buttons as boolean values. The off board and onboard ESP32s are both connected to a WLAN through a secure access point running off of a router. The ESP32 onboard the robot operates as a Telnet server and monitors all communication sent on the server. The TCP packet structure adds 192 bits to each packet sent. This addition increases the minimum packet size to 344 bits per packet and the maximum packet size to 408 bits per packet. The offboard ESP32 Telnet client transmits the sensor information packet at adjustable frequencies. To maximize robot responsiveness and minimize the amount of data transmitted, the packet transmission frequency is set to transmit at 10 packets per second. At this frequency, the communication system operates at data rates ranging from a minimum of 3.44 kbps to a maximum of 4.08 kbps. The robot’s five minute interval of operation will have a total amount of data transmitted over the network between 1032 kb and 1224 kb. 

Onboard Data Methodology:

After data is received wirelessly by the onboard ESP32 Telnet server, the information packet is stored as a string. UART is used as a means of communication between the onboard ESP32 and the Arduino MEGA. Once the Arduino MEGA receives the information, it unpacks the information packet and assigns each delimited value to an Arduino MEGA global variable after every serial interrupt from UART. These values are then cast from their character array form and hexadecimal appearance to their actual hexadecimal value. These values are converted into unsigned eight-bit integers where their values are used to control motors or begin sequences. 

The eight-bit values are used to control the motors due to their level of precision. The Arduino MEGA interfaces with the Talon SRX devices to control the motors through the use of Pulse Width Modulation (PWM). Each Talon SRX is connected to the Arduino MEGA by its own digital pins and logical ground. It can be seen below that the five Talon SRX motor controller is connected to the Arduino MEGA’s digital pins 2 through 6. 

Onboard Power Distribution

The power distribution onboard the robot is designed to supply power to two drive motors, two linear actuators, one excavation motor, an ESP32, and an Arduino MEGA from a single source. The system is powered by a 25.6 V battery. In series with the battery’s terminals are a thermal fuse followed in series by an emergency stop button. Following the E-STOP, the 25.6 V power is converted to 12 V through a high power DC to DC converter. The output of the DC to DC converter is in parallel with a large 75mF capacitor to maintain a stable supply of current when the system is under load. The 12 V leads are attached to a fusebox and the 0 V lead is connected to establish a grounded chassis. 20 A fuses are connected in series leading to the power side of the motor controllers to power the drive motors, linear actuators, and excavation motor. Each Talon SRX motor controller receives a fused 12 V and connection to chassis ground to power the motors. A 1 A fuse is implemented in series leading to a 12 V to 5 V DC-DC converter for the ESP32 and another 1 A fuse is in series leading to a 12 V to 9 V DC-DC converter for the Arduino Mega.The wiring diagram representing both the onboard data connections and the onboard power distribution is shown below.

Software Subsystem

The robot would have utilized an array of cameras, inertial measurement unit (IMU), ROS software framework for localization and path planning. The cameras have 360 degree vision with a 15 degree overlap between fields of view. This allows the robot to track markers placed at key areas and through software produce pose estimates with respect to the markers. The IMU generates information about the orientation and acceleration of the robot. The ROS framework would have a filter to fuse the information of the cameras and the IMU to generate accurate odometry data. The ROS framework would have created a 3D model of the environment as the robot moves around its environment. The model would compute the lowest cost path and send signals to move the robot accordingly. 

The localization of the system was supposed to be composed of four different components. The first component would have been a Logitech USB webcam for the implementation of cameras. The second and third systems are the IMU and an Arduino. The IMU will generate its measurements and the Arduino would have operated the IMU over serial connections. The last component of the localization system was an Intel NUC. The NUC would have received IMU measurements from the Arduino through wireless communication and camera frames. The estimated pose of the robot would have been published to the path planning systems. Controls were established based on the competition requirements and additional implementations of controls were discussed in the electrical system discussion. 

The reliability of the software system is critical to autonomy, position, and safety of the robot. This all depends on accurate measurements and predicted estimates. For autonomy, the use of sensors which generate accurate measurements with low noise. The cameras run 6 FPS, they provide stable and accurate location information. Additionally, the filtering of all measurements could also ensure and provide reliable data for robot safety. 

Next Steps

The robot can be improved in the mechanical, electrical, and software subsystems.

On the mechanical side, the team needs to modify the motor mounting brackets to ensure that the wheels do not interfere. Additionally, it is necessary to mount the excavation arm and excavation wheel. The excavation wheel designed by Team Selene was outsourced to a third party that unfortunately did not complete the job in time.

The electrical side of the robot requires implementing wiring harnesses for the power distribution lines and the signal wires. The electrical system also needs adequate dustproof enclosures that will not interfere with the RF communication system (such as schedule 40 PVC).

The software side of the robot can be improved through implementing high-level commands, autonomy, and on-board camera sensors.

Each of these areas of improvement are documented and going to be a part of the deliverables given to next year’s Lunabotics team.

Meet the Team

Marcus Brown

My name is Marcus Brown and I am graduating this semester with a B.S. in Electrical Engineering with an emphasis in Antennas and Wireless Communications. I am originally from Southern California and came to Mines for the small community environment. In my spare time I like to go hiking and play soccer. My aspirations are to work in the commercial or military aviation industry as an RF engineer. 

Joe Haynes

Joe is a graduating senior in Electrical Engineering. He has interest in embedded systems, RF, and power systems. Joe’s career goal is to work as an electrical engineer in the burgeoning EV industry. In his free time Joe enjoys painting, distance running, and working on cars.

Nadia Scharunovych

Nadia is a senior in mechanical engineering. She is very passionate in robotics and has been involved in robotics projects for four years. Her hobbies include rock climbing, painting, and exploring the Rocky Mountains. Nadia is interested in careers that focus on robotics and additive manufacturing. 

Ryan Weber

Ryan is majoring in Electrical Engineering with a minor in Computer Engineering. His interests include aerospace, embedded systems, and programming. Hobbies include skiing, movies, and weightlifting.

Aneesh Pawar

Aneesh Pawar is from Los Alamos, NM. He is a Mechanical Engineering senior with a Minor in Robotics. He likes to cook and eat food, play soccer, and drink copious amounts of coffee. 

Zeyun Yang

I am studying mechanical engineering and robotics at Mines, and I am planning to pursue future study of robotics in graduate school.

Jason Mackay

My name is Jason Mackay, and I am a senior graduating with my BASc in Mechanical Engineering. I am a member of the Mechanical Team on this years’ Lunabotics design challenge. In this role, I developed the CAD model, aided the design of critical components. Our design required the welding of aluminum box tubing, and so I had the opportunity to learn how to do so. This project has been an enjoyable challenge that has taught us all lessons we will certainly use throughout our careers.

Daniel Abel

My name is Daniel Abel and I’m a sophomore volunteering on the Lunabotics Senior Design team. I’m currently working towards a bachelor’s degree in Mechanical Engineering, with the goal to work in the robotics and/or space industry after graduation. Some of my personal hobbies are designing and tinkering with objects in Solidworks and other CAD programs, 3D printing, and disc golf. I also enjoy playing soccer, and generally getting outside when the weather permits.