ScratchData LogoScratchData
Back to RoboSquid's profile

Getting Unstuck Day 20 - Day 9 revisited

RORoboSquid•Created July 24, 2018
Getting Unstuck Day 20 - Day 9 revisited
5
1
11 views
View on Scratch

Instructions

Getting Unstuck Day 20 - Revisiting Day 9 Create a project that involves the touching color blocks in the Sensing category (i.e., touching color? and color is touching?). Just run the project and the ball will use simple rules to travel around the uncoloured track. It will run over and over with different speed and angle parameters and store them in 3 lists: time for run, speed and angle. Press s/h to show/hide the lists.

Description

I didn't know what to revisit. I spent ages going between projects. I went through the getting unstuck cards and played with some featured projects. When I returned to my getting unstuck studio I saw my "robot line follower" from day 9 and knew it was the one from the notes I'd written about what I'd like to try with more time. I'd never normally write notes like that, so I think I will try to in future to capture my thoughts at the time rather than allowing them to drift away. I have been thinking about line following robots. Seemed like the perfect opportunity to try out very simple rules, eg move forwards; if touching red turn right; if touching blue turn left. Seems to work surprisingly well. I like the idea that Scratch could be used to prototype real world robot algorithms. I started with Scratch Cat rather than a ball, but extremities would "catch" on the colours and lead to twirling and getting stuck. I said: "I'd like to experiment further with different angles and number of steps. Maybe with a timer to see how fast the ball completes the track. This could be automated to determine the best parameters." I've started this. Runs are now automated with randomly selected parameters. I need to sort the list on time to create a proper sense of which parameters work best or even better, plot a surface with the 2 parameters to see the "shape" of the parameter space. I also need to add jitter to the start point to see how robust each result is. Sometimes the robot cheats and returns to the gate by going the wrong way down the track. Maybe need multiple gates to track this? Oh dear - I said I was going to get an early night tonight...

Project Details

Project ID237300448
CreatedJuly 24, 2018
Last ModifiedJuly 25, 2018
SharedJuly 24, 2018
Visibilityvisible
CommentsAllowed