There were no major changes to the existing design.
Challenges & Mitigation:
- Datasets for ML not having all necessary data: There were no single datasets that have all the data we needed for health detection, so multiple datasets were selected to be combined.
- Potential plant loss during training: We planned to let the system support three plants last time. In order to ensure a larger dataset and be ready for plant loss, we ordered 4 plants for each plant type. We are planning to set up the sensors and actuators as quickly as possible.
- Increased costs: We ordered plants and additional components (blackout window film, liquid nutrients, soil, basil plants, hamalayamix foliage plants, flowering plants), which was a larger increase in costs than we expected. To mitigate this, we found the cheap but appropriate components to be purchased and minimized the amount of extra components like soil.
Progress:
- Finalized and ordered plants and additional components for growing plants.
- Assembled the greenhouse.
- Worked on code that runs sensors.
- Started working on the design review report. Split up sections of the report for each person.
- Keep working on ML integration and web app.
Next Steps:
- Keep setting up temperature/humidity sensors and start setting up a heater.
- Finalize web app and user interface.
- Establish initial training framework for the ML model.
- Keep working on the design report.