MIME 4450 Lec 15 10 29 26
Overview
The lecture connects agricultural automation to hands-on robotics projects: a vision-guided system could select propagation leaves for Ball FloraPlant, while a class project already links a PLC, robot arm, camera, and HMI to play a dice game. Students are asked to choose project hardware and document their effort, with guidance on Siemens PLC communication, Hiwonder setup, Ubuntu and Python, and the practical need to replace industrial controls every 6 to 8 years.
Key takeaways
- Automating Ball FloraPlant’s propagation process requires solving the vision task of selecting harvest-ready leaves; the robot arm and downstream potting equipment are not the main technical bottlenecks.
- A labeled dataset of human leaf selections could train an AI vision system to distinguish suitable cuttings from leaves that should stay on the stock plant.
- The demonstrated dice-game project integrates a PLC, robot arm, camera, and HMI, with Ethernet connectivity required for the robot to receive commands and return results.
- For course projects, students must select hardware and work in teams of no more than three; Hiwonder projects require basic Ubuntu and Python familiarity, while FANUC offers practice with user planes and tool frames.
- Industrial automation work includes maintaining and replacing control hardware: the lecture gives typical PLC replacement intervals of 6 to 8 years and notes that 2011 lab processors were failing by 2026.
- Project assessment prioritizes documented effort and troubleshooting, and permits alternatives such as simulated conveyor movement when a physical conveyor is unavailable.
Chapters
- A stock plant supplies leaves or cuttings that are rooted in soil to produce new plants.
- Ball FloraPlant harvests cuttings in Nicaragua and ships them to the United States for potting, serving greenhouse markets including California and northwestern Ohio.
- The proposed project targets the unautomated step: finding a suitable leaf on a stock plant and cutting it.
- Nicaraguan harvesters select and cut leaves by hand, with crop-specific cutting standards and a target of 105 cuttings per bag.
- The facility shown uses controlled greenhouse conditions, including automated vents and fans, crop-specific fertilizer schedules, and recycled irrigation water.
- A separate U.S. machine already transfers cuttings into pots, but the upstream task of selecting leaves from stock plants remains unresolved.
- The difficult capability is 3D vision that distinguishes cuttings worth harvesting from leaves that should remain on the plant, not the robotic arm itself.
- A training set of human selections could label good and bad leaves so an AI model learns subtle selection criteria and directs the arm.
- Cognex vision systems were described as progressing toward this sophistication, while the camera on the class’s Ubuntu-based robot was said to support the capability.
- Automated harvesting could bring production to U.S. growers and reduce shipping between Nicaragua and the United States; related research includes robotic grape, cherry, and apple picking.
- Two classmates built a system connecting a PLC, robot arm, camera, and HMI to operate and score a dice game.
- The robot retrieves dice and moves them to a scoring area; the HMI lets players start, reroll, or bank points.
- The demonstration explains scoring, including 100 points for ones and 50 for fives, and shows the interface responding to game outcomes.
- Disconnecting the robot’s Ethernet link makes it unreachable and prevents commands from being sent, demonstrating the system’s dependence on its connected components.
- The instructor shared a NoMachine network configuration: an address beginning 192.168.2 and subnet mask 255.255.255.0; the complete host address was not stated clearly.
- A classmate project supplied PLC communication hints, and Professor Robert Langdorfer’s JetArm guide was promised by email and on the course site.
- Available class equipment included 12 Hiwonder units, 14 Cognex vision systems, safety PLCs, and approximately seven industrial robots after correcting an earlier count of eight.
- Students reported interest in five Hiwonder projects, three yellow FANUC robots, one white FANUC robot, and two Universal Robots.
- For Universal Robots, the recommendation was to use the newer 5e model; sharing may be necessary when equipment demand exceeds availability.
- Teams may have one to three members, must identify members and desired hardware, and should avoid situations where one person does all the programming.
- FANUC’s useful advanced feature is setting offset user planes and tool frames, a coordinate-system skill relevant to industrial robot work.
- Hiwonder projects require some Python and Ubuntu, both already installed on the robot; the course will provide a Python text but will not teach the language from scratch.
- Ubuntu is a Linux distribution, and students are expected to learn enough of its environment to operate the robot and its more advanced vision tools.
- The instructor reported that PLC equipment commonly needs replacement after roughly 6 to 8 years, while some manufacturing equipment may last 7 to 10 years.
- EET 4450’s ball-in-a-tube lab used processors purchased in 2011 that began failing outputs by spring 2026.
- Updating the labs exposed Siemens software and certificate complications, illustrating the real-world work of keeping control systems operational across equipment generations.
- Students may build a dice game, another automation task, or a different project; a real conveyor is optional because a simulated part movement is acceptable.
- The grading emphasis is on sustained effort and documentation of work completed and problems encountered, even if the project is not finished.
- Assignment 8 was due at the end of class, and Assignment 7 was also accepted from students who had not submitted it earlier.
Transcript excerpt
0:06 So, um, this is a stock plant and it would be a plant that you would use, uh, to basically cut leaves off of in order to make other plants. So, this would be one of these where you'd snip off a leaf and then put the leaf the the stem in the ground and then it would grow another plant. And there are a lot of plants that are that are made this way. Okay, this is a stock plant and there's a lady over in Wolf Hall, Jennifer Bolt, works with the USDA that has some of these and she introduced this project to told me about it. She said they've not been able to do this in the United States. What they do is they basically
0:53 get the stock plants in uh Nicaragua and then they cut the the the the uh leaves off and then they ship them to the United States and then they put them in pots here in the United States. And this is mainly a California based industry, but it is also done in in northwestern Ohio because the uh of the of the large number of green houses in this area. And I've actually seen the robot that does the the back end of this, which is the I'll show that to you in a minute. But there is no robot presently that actually goes in and finds the leaf in this plant to cut. All right. So the goal of this project would be to find the leaf on this plant that is ready to
1:42 be cut and then eventually taken and then repotted. Okay. So this is the half of the project that is still out there in the wind. So that's that's that's what this is about. So anyway, this is a
This is the start of the transcript. Save this video free for the whole transcript, searchable and time-coded.
In this course
Related notes
Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Wm_T Evans.