Monday, 15 February 2016

Sprint 1 reflections - Joe Tubbritt

Upon starting the first sprint, the team was assigned the following tasks: to combine servo motor control, sound generation for movement of servos and face detection via OpenCV, upon the Raspberry Pi. The tasks were divided up into 3, one for each member of the team: I was to write code for the servos, Willy was to do the sound and Ian, the face detection. Once these tasks were completed, it was required that each task be combined in a thread. The idea being that when a face was detected the servo's would move to "track" the face, and sound would play while movement was occuring!

Technical reflection:

Upon completion of the first sprint, we were unable to fully integrate all the tasks into a working thread. However, it must be noted that while we could not demonstrate the combined task thread, we did manage to get the individual tasks to operate independently. For example, my servo code worked like I had intended it to. Willy had sound playing, and Ian had the camera detecting a face via OpenCV.

Looking back on the sprint from a technical point of view, it must be said that the servo's require an external power supply, rather than having them run off the Raspberry Pi. Servo's draw a lot of current and this caused them to become "jittery" when stationary. This random movement stems from the fact that the Pi cannot fully handle the current requirements that the servo's need. Also another point which I would like to highlight is possibly using servo's with more torque. I noticed during testing that sometimes the servo could not lift the metal casing surrounding it (tilt servo). It would struggle for a  bit and then eventually lift it. Using a higher torque servo (current servo is rated 13oz-inch) would improve movement and overcome this problem.

Project management:

In terms of project management, our team is unusual in that we still haven't decided on a team name yet (We are not being called Team Joe guys!). Other than that, everyone is able to manage their time suitably to accomplish the tasks we were designated. Our group does not have a leader, but I feel that everyone is pulling their weight and we are getting stuff done (bar a few problems along the way). It was disappointing that our thread did not work but I feel that this was not due to bad project management, but rather a combination of bad communication and bad luck!

Personal reflection:

I personally found the first sprint to be enjoyable. I have prior experience using servo's so I knew I could deliver the task. It was new ground for me to use Python, as I had never programmed in it before. It was an interesting learning experience getting to grips with Python basics and learning how to program the servos. Also, it was new ground using the Raspberry Pi. I had seen some of the lads over the years using it, but I unfortunately never had the chance until now. I can see why everyone used to highly recommend them. Usng it is fun and enjoyable. I like how it is very much a mini PC!

This project has the potential to be highly interesting, depending on the direction we all take. I would hope that we can reach Jason's expectations, surpass them and produce something entirely unique and well designed.

Teamwork reflection:

In terms of teamwork, I think our team lacked it the most during this sprint. I feel that we did not communicate enough with each other. In the first week we kind of did our own thing rather than conferring on a plan of action. However, as it became closer to the deadline for the sprint, I do feel that we started to become more comfortable with each other and we discussed the project a lot more. To be a good team, we have to communicate more and contribute ideas. I felt from watching the other teams that we did not have much creative ideas to explore.
In fairness to the lads, we have come together since and are stating to get our asses in gear with the project, and I am enjoying being a part of our unnamed team!

Next sprint:

For the next sprint, I definitely think we should get lamps in, so that we can start to figure out things like: getting servos mounted for movement, how will the Pi be connected, how versatile is the design i.e. movement ranges etc. To see something moving and tracking someones face while playing sound would be the final goal of this sprint. Personally, I think this sprint should accomplish that quite easily.

Joe - Blog #3

No comments:

Post a Comment