Navigation and Optimized Support Rover

During this project, we will design and manufacture a service and maintenance rover for a long-term mission to establish a permanent lunar base. Mechanically, the rover will be designed with characteristics that allow it to operate for extended periods (~1 year) without requiring in-depth repairs.To achieve this goal, we will use cycloidal systems for the movement of mechanical components, such as the 6-degree-of-freedom (DOF) arm, due to their low backlash compared to traditional spur gear reductions. Other components that will contribute to this objective include permanently lubricated bearings, as well as aerogel and multi-layer insulation (MLI), which will significantly reduce the amount of supplies the rover requires over time. As a direct result, this will decrease the long-term operational costs of the rover.The rover will have a mass of approximately 100 kg and dimensions of 1 m × 0.6 m × 0.3 m. The 6-DOF arm will be capable of lifting loads of about 50 kg, which corresponds to 81 N (18.2 lb) under the Moon’s gravity (1.62 m/s²).For materials, we plan to use the 7xxx series aluminum due to its excellent mechanical and thermal properties relative to its density Our lunar rover will be equipped with an extensive array of advanced sensors and components to ensure functionality under lunar conditions. The main brain of our system will operate through the Jetson Nano, allowing our team to seamlessly integrate data from modules found throughout the rover. These sensors will include but not limited to the Zed Mini Camera, multiple Micro Lidar sensors, and IMU modules. These will allow our rover to have precise navigation and multiple layers of obstacle avoidance powered by our autonomous navigation model. These systems will be powered by a Lithium Iron Phosphate battery similar to the BYD’s Blade battery technology. This allows for unparalleled safety, durability, and lifespan when compared to the traditional lithium-ion batteries currently implemented. This also encourages hot-swapping of the batteries so that the rover can have extended periods of activity. Together, these components will allow our rover to complete intricate tasks with accuracy and dependability. Using the information provided by our sensors, we intend to have autonomous control for both navigation in predefined areas and for manipulating the payloads mentioned before. We will incorporate data from a PID controller built into our embedded systems to automatically make calibrations, allowing for more dynamic movements to account for a changing environment. The training of the computer vision system will be based on a convolutional neural network system tailored for different live input from our sensors. We intend to also experiment with modeling training using Gazebo software to model how output from our model affects behavior of our mechanical systems.