Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 4 | Agricultural Drone Refilling System |
Aditi Adya Batu Palanduz Steffi Chen |
Yixuan Wang | design_document1.pdf final_paper1.pdf photo1.png photo2.png presentation1.pdf proposal1.pdf |
|
| # Team Members - Batu Palanduz (batup2) - Aditi Adya (aditiaa2) - Steffi Chen (steffic2) # Problem With many agricultural drones, the sprayer tank needs to be manually refilled rather than having an automated system. While this does not pose a problem if there are a small number of drones, as the fleet size increases, tank refilling will take up more and more time, questioning the efficiency of this current system. This will result in a decrease in productivity as more time will be spent refilling the tanks instead of operating the drones or taking care of other tasks, such as analyzing the data collected from the drones and performing maintenance on various equipment to give a few examples. # Solution Overview An automated refilling system would relieve this issue by refilling the empty sprayer tanks without human intervention. This would free up the farmer and enable the drone fleet to operate more efficiently by reducing the downtime caused by waiting for an empty tank to be refilled. The refilling system would consist of a gantry that contains the refilling nozzle, camera, distance sensor, and pumping hardware needed to align the nozzle to the fill port on the drone's tank and refill it. Additionally, a computer and microprocessor would be needed to handle the image processing from the camera and control the gantry motors, respectively. Visual markers can be used to determine the location of the fill port, as well as the distance to the fill port, using image processing. The distance sensor would act as a backup to ensure that the gantry does not accidentally crash into the drone if the image processing fails to correctly determine the distance to the drone. # Solution Systems **Refilling System** - Tank Subsystem - Has a fluid monitor which signals to the control system if the refilling station needs refilling - Dispensing Subsystem - Has a distance sensor, nozzle, and hose which handles delivering the fluid to the drone - Gantry Subsystem - Uses stepper motors to move the dispensing subsystem in a controlled and precise manner. Has stepper motor drivers to power the stepper motors - Computer Vision System - Uses a Raspberry Pi for image processing and a camera for accurately aligning the dispensing subsystem with the drone’s fill port - Control Subsystem - Controls gantry movement and monitors the refilling process to prevent drone overfilling. Also monitors the tank subsystem’s fluid level and displays a notification if the tank needs refilling **Drone Replica** - Represents a replica of the important parts of the drone: wing/fuselage area around the fill port, fill port, visual markers, tank with fluid level sensor, refill the status display **Power System** - Includes an AC/DC power supply and off-shelf voltage regulator(s) to provide the needed voltages for the subsystems # Criterion for Success Our solution will be able to accurately refill water into the tanks of the drones. The detailed criterion for success is as follows: Precisely recognize the entry port to the water tank and line up to the tank port Make sure there is minimal to no amount of extra spillage around the water tank while connecting, filling, and disconnecting Correctly sense when the tank is filling up so that the refilling system does not overfill it or stop at the wrong time Send a signal to the drone to show that it is done being refilled # Anticipated Difficulties Some of the anticipated difficulties revolve around the integration between the hardware and software aspects of the project. Troubleshooting and debugging the gantry movement and alignment will take a long time as there are many sources of error that need to be accounted for, including slop in the mechanical system, repeatability, and any design oversights/errors. Difficulties with the software aspect might include difficulties reliably identifying the visual markers in different lighting conditions, dirt or other debris obstructing the visual markers, potentially steep learning curves to image processing/recognition, and reducing the computational power required to minimize costs. |
|||||