Week 1 (9/4 - 9/8)

Milestones:

• Met with the project sponsor/client to ascertain the scope and required result of the project.
• Assigned roles to team memebers.
• Played around with Darknet YOLO object detection and began discussing preliminary architecture.

Conflicts:

• None.

Outstanding Issues:

• None.
  

Week 3 (9/23 - 9/29)

Milestones:

• Created a way to collect training data and started on detector commands.
• Searched for more viable cameras.
• Began development on the pre-processing device.
• Continued exploration of firebase, tensorflow, and networking solutions

Conflicts:

• React Native running and how to synchronize iOS and Android.

Outstanding Issues:

• Still unsure of networking solution.
• Mounting options are still up for debate

Week 5 (10/7 - 10/13)

Milestones:

• Design document completed.
• Finished up development of layers for ML model.

Conflicts:

• Still need to find a permanent solution to our Python setup environment.
• Running into out of memory error when I try and run the learning algorithm on a tensor.
• Getting runtime errors when trying to get spot information and represent that data in a React Native app.
• Issues with trying to figure out the Python firebase library.

Outstanding Issues:

• None.

Week 7 (10/21 - 10/27)

Milestones:

• Finished basic functionality of app.
• Finished error reporting and logging functionality for pre-processing device.

Conflicts:

• Compiling Tensorflow to be used with GPU. As well as work on real-time predictions using the trained model that we are developing.
• Lots of issues with compiling sources of tensorflow.
• Change app code from using Firebase’s Realtime Database to Firebase’s Cloud Firestore.

Outstanding Issues:

• None.

Week 9 (11/4 - 11/10)

Milestones:

• Began real-world testing of our system.
• Running tensorflow on GPU.
• Finally got React Native and Cloud Firestore to work together.
• Pre Processing device setup.

Conflicts:

• List Items are not being created with the spot information gathered from Cloud Firestore.
• Getting started on react native with Joe and John and get some of the slide designs actually implemented.
• Finish the queueing of parking spot images during predictions.

Outstanding Issues:

• Getting Cloud Firestore to compile and run with React Native.

Week 11 (1/8 - 1/21)

Week 13 (2/8 - 2/22)

Week 15 (3/9 - 3/22)

Week 17 (4/7 - 4/20)

Week 2 (9/9 - 9/15)

Milestones:

• Finalized dates and basic tasks in order to create overall project timeline.
• Began to do research on Tensorflow.
• Reaserched possible hardware.
• Began planning possible camera placements.
• Started mobile app template.

Conflicts:

• None.

Outstanding Issues:

• Finding color schema and name.
• Waiting on the OK to use Firebase.

Week 4 (9/30 - 10/6)

Milestones:

• Set up enviroment for easy deployment.
• Wrote script to get Raspberry Pi serial number and update Firebase.
• Starting setting up mobile/web enviroments.
• Created issues on github to start development.

Conflicts:

• None.

Outstanding Issues:

• Post-processing training is going to take a while. Need more issues on the gitlab for iOS, Android, and Web.
• Keeping the different applications the same visually. (iOS, Android, Web).
• Creation of our Machine Learning model using Tensorflow

Week 6 (10/14 - 10/20)

Milestones:

• Began to set up the docker container.
• Finished rough training method. Got to 77% accuracy.
• Created a basic app in React Native.
• Completed image slicing and script to move documents between pre-post processing.
• Created an error handling email that will receive notifications of any pre-processing issues.

Conflicts:

• Finishing the predictor model and the Docker container.
• Training time takes a long time.
• Learn React Native and integrate it with Google Firebase.

Outstanding Issues:

• None.

Week 8 (10/28 - 11/3)

Milestones:

• Finished implementation of the predictor.
• Reviewed and managed merge requests, created documentation for build of tensorflow.
• Have app connected to Cloud Firestore.
• Began screen mockup to add to the next version of the design document.

Conflicts:

• Try to get the model to train using the GPU, as well as look into ways in order to optimize our current model.
• Finding a machine that can run the tensorflow gpu.
• Getting Cloud Firestore to compile and run with React Native.

Outstanding Issues:

• Getting Cloud Firestore to compile and run with React Native.

Week 10 (11/11 - 11/26)

Milestones:

• Finished implementation of the predictor.
• Reviewed and managed merge requests, created documentation for build of tensorflow.
• Have app connected to Cloud Firestore.
• Began screen mockup to add to the next version of the design document.

Conflicts:

• Further work was done to try and integrate firebase with react native, but we may need to start looking for alternatives.
• Finding where our model is failing only getting 50% accuracy for an unknown reason. Going to look into our model and see why.
• Pushing files from Heroku to gitlab.

Outstanding Issues:

• Getting Cloud Firestore to compile and run with React Native.

Week 12 (1/22 - 2/7)

Week 14 (2/23 - 3/8)

Week 16 (3/23 - 4/6)