Sunday, 8 February 2015

Sprint One - Reflection - Seán Doyle

The Team

The team was comprised of David, Hervé, Peter and myself.As a unit we worked without any conflicts arising between ourselves and I believe we got along well. However within our own group we spent the majority of the sprint split into two sub groups with David and Hervé in one who were primarily working on mutex's and threads and Peter and I working in another team figuring out the basics of the hardware and formulating some ideas for it. In hindsight I still think that splitting into two sub groups was the right decision (as having 4 people trying to operate off of one small device would be very awkward) however our communication between the two groups would need improving. The result of this shortage of communication could be seen when we finally combined the two groups work, with this team based project being a new experience (at least for me) I will strive to improve communication from myself onto others and vice versa.


From a technical perspective I feel that our end result of the sprints was quite good with it successfully tracking an object on the horizontal plane in front of it. The design resulted in a face like appearance and operated using two sonar's to track the object.
Also having never worked with any hardware before I found this to be an enjoyable experience learning about the Servo's and Sonar's.

Moving Forward

For now I would like to continue working on the system that our team designed using sonar tracking but with a camera implemented so that we can add shape/object detection. After watching the demonstration on the camera face/ball tracking I found it to be very slow looking and for the lack of a better word 'clunky', as such my suggestion would be to look into minimising the amount of camera use and try relying on a faster sensor such as the sonar. The camera would be used on a fixed interval or on a triggered event (ie. it suddenly loses its target) to allow the lamp to detect a person face and then we would proceed to using the sonar's to track said face. As the main goal is to replicate the Luxo Jr. scene of the ball rolling by (shown below in video) and based off of the impression I got from the demonstration I believe the camera wont suffice as the primary tool.

(0:22 into video for the relevant scene)

Alongside the above suggestion I also believe it would be a good idea to get a proper end design sorted out so we can plan for that accordingly rather that prototyping ideas that may not work in the end product at all. 

-- Seán Doyle

No comments:

Post a Comment