Be sure to listen to this week's update on CASE Wednesday!

 

The second week of the project started with the group splitting into a Software and Hardware team. The teams were: Andrei, Alex, Aye Mon, Hsiang Hsiang on the Hardware team and Sam, Jason, and Jurgen on the Software team with Josh a part of both teams. Once the two teams were formed, each team determined the deliverables for the week. After this was done both teams set out to do their respective parts.

 

The Hardware team started by discussing possible methods for turning the dials and pushing the buttons. Initially, the team thought of a pulley system for turning the dials, however over time this could lose its accuracy. For pressing buttons, the team thought of potentially using actuators. However, it would be a challenge to correctly provide a consistent pressure every time a button was pressed. The team quickly realised that servo motors would be a good choice for both tasks. Thus with the components determined, ideas of how to make effective use of the motors to save space for the final case were generated. After some thought, the team agreed that the best approach would be to remove the current dials, and 3D print components that can be attached to the servo motors and do the same job.

 

With LEDs on some buttons which need to be taken into account, Hsiang Hsiang and Aye Mon worked with Josh to design a part that could attach to the servo motor and also hold an LDR and allow for the buttons to still be pressed. Alex and Andrei also worked with Josh to design a new component that would fit into the spot of the dial but connect to the servo motor. Once the designs were confirmed, Andrei went a got both 3D printed for quick testing.

Alex was developing the code for the servo motors so they could be used with the Arduino. By the end of the week, the Hardware team had developed multiple designs for the 3D printed parts, and had most functions using the Arduino working individually. This included the servo turning the dial a certain amount based on an input, the LDR determining if the LED on the coffee machine was on or off, the servo motor moving left or right so that one motor could press two buttons.

 

The Software team started by getting acquainted with NodeRed and Bluemix, the different ways of using them, and how it could all be run using the Raspberry Pi. The first task that the team did was to get the Raspberry Pi running correctly. This included being able to connect to the internet via WiFi, and installing NodeRed to allow for easy implementation of the Nodes that were made in NodeRed to run on the Raspberry Pi, and allowing for the Raspberry Pi to be used with a touchscreen. Jason and Josh worked on getting this done.

 

The Software team divided tasks into 3 main categories. These were: Facial recognition, speech detection, and creating a dialogue tree and chatbot. Sam worked on getting the chatbot and dialogue tree working. By the end of the week, he managed to use a form on the project website to provide an input on NodeRed and then for the user to receive an email notification that their order had been received. He along with Hsiang Hsiang managed to get a dialogue tree for ordering coffee, its strength and how many shots running.

Jason worked on getting facial recognition working and linking this to NodeRed. Using the Watson API for facial recognition and a few pictures of different group members, by the end of the week, he had managed to use a camera to get a picture of a group member and the member would be correctly identified. Jurgen worked on speech detection using the Watson API's and NodeRed. He managed to get a flow on NodeRed that would take in a voice recording, and would correctly turn it to text and then take that text and repeat it back to the user.

 

On top of working on the hardware and software aspects of the project, a website was created to follow the group's progress. Andrei is working on the website and providing updates to how the group is progressing. To provide more information, a podcast will accompany the weekly updates. Aye Mon works on the podcasts providing a nice way for anyone interested in the project to follow the groups progress.

 

After the first full week of working on the project, the group has made large strides in the right direction. The Hardware team has gotten most of the individual functions needed to work, while the Software team has managed to make good progress on facial recognition, speech recognition and creating a chatbot that can be used in the final product. The following weeks will consist of refining what has been completed to make it run as well as possible and connecting the hardware and software together. This will involve determining a viable way of connecting the Arduino to the Raspberry Pi. After this is done, a sleek case must be designed which allows for all the hardware to be intricately connected to the coffee machine and create a 'SMART' coffee machine!