Team: PI over Four
Author: David O Mahony
I feel like we have met the basic goals and are just shy of having the full working system were the camera sets the value of x y and z with both sound and motors acting on it. I am proud of the work that team has achieved and look forward to working further with them on the project.
Areas to improve:
· A lot of the work for this sprint was done on different raspberry pis and then placed on to Phil’s own board. It is my thinking that from now on when people on the team get things working that it is then installed on a set aside pi that will have only working code this means that if people are out or missing that others on the team will still have a platform that is fully operational and can there for test and keep working on improvements.
· I think it is also important to get a second XMOS starter kit board. This means that the set up can always remain in the college for people to work with and to test code with motors.
· I think it’s important that we get lamps soon as it will help motivate teams and to progress the project as teams can start to see how the lamp will move with their motors.
· I think it is important for us to improve are threading and look in to how other teams have implemented there’s which react faster than are current set up.
Ideas for next sprint
· An Idea I would like to implement would be to bring in telepresence aspect to the lamps. Such that a user can control the lamps movement over the internet and view a camera feed as well. This set up would only need camera, movement, sound (not necessary) and control over the web.
· The set up of idle routines such as moving in a set pattern and a set sound.
· Test and implement controls for a light bulb, could be simply implemented through the use of a optocoupler to allow mains voltage to turn on a bulb while only using the normal pin powers on the board.
I feel as a team we worked well but can improve how we can communicate to each other and should organise more time to meat up outside of the class times. I think that if we had managed to meat up more in the last week when we were combining everything we would have managed to get a full system working. An idea put forward in conversation with my class is that the main teams should stay the same but with extra teams created within the class they will contain the members from the other teams that deal with specific tasks. Such as a team that will work on sound and a team for vision. This will improve team to team collaboration and increase the pool of knowledge in the class.
From demonstrating the project and looking at other groups attempts there is quite a lot of work that needs to be done to improve the threading of the project in speed and efficiency. Are current set up is sluggish and can be buggy at times. I think if we can converse with other teams we can improve are set up. The camera system worked well but due to difficulty reading faces it may be of benefit to change to track something like a ball which is the same from all angles unlike a face. The sound works fine for a demo setup but I see that in the next sprint I would like the lamp to play a different sound as it moves away from the center. To do this I would break a sound file in to 1ms clips that when played to gather plays a full sounds. Thus as the lamp moves from the mid position out the sound will change as it moves. This means that the sound will change quicker and sync up better with movement as the sound clips can be incredibly short.
The movement of motors seems to be good as Filipe system will be able to control multiple servos which will need when we start controlling the full lampbotics. I feel like we should push this in testing and control the full number of servos that will be needed in the final sprint and see how the board handles the test.
David O Mahony entry 4,