This is the final week for getting things done and submitted so we have scheduled in an extra session of the club.
The TT motors are playing up and giving very little power on one side which is making steering difficult. We swapped 2 of them over so that it was more evenly powered on each side as a temporary measure then urgently bought and installed 2 new motors.
This week we need to run and video the obstacle course and feed the fish challenges
Further tests on the ball flinger show that it can only throw 3 balls far enough on the course so we will do the 3 runs with 3 balls.
Here is a picture of it throwing 3 balls successfully into the top part of the container
When we videoed it for the challenge we got it to do three consecutive runs, each throwing 3 balls into the top container – making 9 items in the box in all in 59 seconds.
We then made and ran the obstacle course – using spanners to spell out the word PI WARS as the first part of the course to be run over.
This was followed by other obstacles, including a see-saw and ramps
We videoed this .
Our only worry is that because of the battery problems which were blowing up the Pi we had to run the challenges with an external power supply for the raspberry pi. We hope that we will not be disqualified for this after all the hard work that we have put into this.
We had planned to have done all of the challenges autonomously and we even had a remote voice control unit working, but with the repeated equipment failures and the very limited time we had after release from Covid-19 lockdown we ended up doing them all using remote control from a gamepad.
We are having problems with one of the line following sensors for up the garden path, so as a fall-back we have got the remote control working to be able to steer the robot around the course
Here is the robot at the start of the line follower course. Because of the ongoing issues with powering the pi from the battery which has now destroyed 2 raspberry pi 4bs, the pi is being powered by an external 5v supply, but the robot motors are being powered from the battery
We also ran the “tidy the toys challenge” and got a video of that.
Here is a shot of kesbot3 at the start of that challenge:
Having problems with the raspberry pi when used with the 3 cell lipo battery. The Adeept motor hat which regulates the voltage down for the pi has a chip rated for up to 30 volts, but when we put 12 volts on it the pi stops working and will not reboot. Looks like the pi has been blown.
Unable to do much more this week until we get new Pi
Open CV code is working with the pi camera and we can now detect blocks and their position on the screen. We now need to use this information to link it to the movement code to take the robot towards the colour block needed.
We did some more testing of the ball throwing device. It should be able to reach the top of the feed the fish container.
We got a new member today, Advait. This means that we now have 7 members! Dhruv showed him what we have been doing on the Raspberry Pi code
During this session, James, Aarav and Marin 3D printed multiple balls to be used in the Feed the Fish Challenge. We print the balls as 2 hemispheres then stick them together.
Marin also saved one of the 3D printers from malfunctioning by removing some of the gunk on the edge.
James, Pavle and Advait (the new member) assembled the 3D printed parts of the 5 ball throwing device for the feed the fish challenge. Pavle sawed and put together the wooden parts for the attachment of the ball thrower to the robot, and drilled some holes to accommodate the screws. This is how it looks: The lead is for the servo that triggers the release of the balls.
Akshat and Dhruv wirelessly transmitted audio commands from one hand held Pi to another one on the robot, which was turned into text using the module server. This will be used for the voice controlled commands for the line follower.
Dhruv also finished the block picking code for the tidy the toys challenge.
This week we have been working on the projectile feature of the robot, as the original idea of firing nerf bullets is shooting them too fast and is unable to get them to drop into the top of the tank because the angle is wrong..
So, 2 members of the group spent their time designing the system of which projectiles could be flung off the robot into the Feed the Fish tank made last week. They started designing this concept based on the following loose design of a trebuchet- a medieval throwing machine.
Based on this Pavle began working on making a trebuchet like design to throw projectiles. He started designing this on CAD.
Another member worked on designing the projectiles that would be launched. He used Autodesk fusion to create two hemispheres as seen below, which can be glyed together after being 3D printed
Pavle also worked on laser cutting the sensor holder after designing it on CAD- here is the image of this being cut:
We 3D printed some of the projectile hemispheres as shown below:
Marin worked on printing the 2cm diameter circular projectiles from the Autodesk CAD design and then glued them together to make this :
Dhruv has been working on the code for the positioning of the robot for shooting the projectile balls designed earlier:
He also worked on the servo trigger for the shooting of the projectiles. Here is his code:
Akshat worked on fixing errors with the remote control but still has somethings to fix.
Robot club is on Easter holiday for 3 weeks now and will be back on 26th April
This week Pavle worked on assembling and gluing together the feed the fish obstacle course using the CAD model developed by James as guidance.
He also used tape to mark the robot’s position and we tested where the bullets were going to shoot, however currently they are shooting too hard, so we will look at a solution based on a Trebuchet instead.
Aarav has produced a CAD model for the 2cm diameter balls that the trebuchet will fire. We will 3D print them on the school 3D printers. Marin has been working on the text for the blog.
Dhruv and Akshat have been working on the infrared sensor programming and have fitted a new camera to the robot after the previous one failed last week. They have also been working on setting up the mic to control the robot. Akshat tried to do this on a pi zero but he could not get it to run so he has taken an old pi 1b home to try on that.