Comparison of project output against original objectives

• WP1 - Sensors: Different options were evaluated for the measurement of nitrogen levels and a Soil NPK sensor was decided as the best option. This sensor also allowed for the measurement of phosphorus and potassium which are important metrics in soil health. The NPK sensor and the DHT11 (temperature and humidity) sensor were tested (NPK sensor by testing in soil and fruit, temperature sensor with room thermostat and stream). Unfortunately, it was difficult to determine whether the NPK sensor was working because it outputted values that looked quite random and the only way to test if these values were right was purchase another NPK sensor which the budget did not allow. The hardware simulation shows that the method for using the sensors was correct. The NPK sensor and DHT11 were next added to the WP2 which is another goal for this work package. This means that all the objectives for WP1 were completed but there was no method to perform accurate testing for the NPK sensor.
• WP2 - Central Hub: The final central hub design was able to accomplish all the required functionalities it needed as well as prove proof of concepts for all stretch targets. The design is also as small, low power, and low cost as required for this product. The hub is an affordable, short-distance communication device that can be used with no prior infrastructure requirements. The software is currently designed to work with the sensors to display and process soil data but it is feasible to redesign the software for a different purpose without having to change the hardware at all. The device is very versatile and therefore could be used with a wide range of different sensors and devices, with just small changes to the software. The next steps for the central hub would be an improvement of the software, adding more features, and making it more robust. In conclusion, the central hub was able to meet all expectations and more and the final product is a flexible, effective product that can be used for even more than it was built for.
• WP3 - Machine Learning: Overall, the objectives of this work package were met well, as the decision tree model provided a very high validation accuracy of 98% and all (most? - TBC) the test cases were accurately classified. Once the model was trained and tested in MATLAB, C code was generated for the model so that the code could be compiled into the microcontroller. After some initial issues with file attachments, the code was successfully uploaded to the microcontroller and produced accurate decisions from the model based on the sensor readings and based on some test cases. This code was also implemented into a website app containing an online database of the sensor readings for further information for the farmer, meaning that one of our stretch targets was also achieved from this work package.
• Entire project:

Evaluation of project time management

The project was managed well, as our progress was tracked weekly through our Gantt chart and this was updated when there were challenges or time delays. The meeting minutes were completed every week and actions for each week were determined during the meetings, so that we could compare and make progress from the previous weeks. Our actions were updated and tracked on Notion, on the Team Tasks page as shown below, including the deadlines for these actions. Each member of the group was assigned a task.

Team Tasks

We also updated the ‘General Progress’ pages every week, so that we could track our progress against the Gantt chart and project objectives. This meant that we could also show proof of progress over each week to our supervisor, and each member of the group can cross reference against each work package to check other members’ progress per week.

General Progress - Term 1

General Progress - Term 2

Finally, if there were any time management issues, such as other assignments, this was communicated to the group and each member was notified that progress would be slower for that specific week. This was also tracked on our ‘General Progress’ pages. If there were issues of any electronics or software not working, we attempted to fix this as soon as possible and updated our timelines on the Gantt chart. We had one challenge of the mechanical aspect of the project not being completed by the leader of the work package, so we communicated that to our supervisor. We overcame this challenge by agreeing as a group that other members of the group would create a simple mechanical design once all our objectives were complete. We had completed all our objectives early, meaning that we had time to complete this mechanical aspect of the project. Overall, our time was managed well and our progress was constantly tracked each week.

Evaluation against professional engineering competencies

Evaluation of working as a team/communication

All progress and updates to electronics or software was communicated to the group via our WhatsApp group chat or through meetings with/without our supervisor. We met once a week with our supervisor, and at least once a week individually as a group to discuss updates of each work package and the next steps for the week.

(Anything else we should add??):