August & Sander: pictures and an expression of what this week means to us:

Where to start?

Lets just start by saying that this is how our project, our Astrorover, looks like after integration hell:

When writing this, after all the hours it took putting the frame together, designing it, gluing it, mounting the sensors and cameras, taping and sorting all the cables inside and connect them to their respective pins, drilling and mounting the weels and motors, without even mentioning all of the hours, iterations, discussion and implementations to actually see our envision standing in front of us, physically, in real life, we feel proud.

We didnt quit. We have put in the hours, and been trusting the process. And here we are. Of course, we are not done. We will iterate everything at least one more time, but the thing is, we dont want to do anything else. The car is a result of our frickin minds. The system as a whole, including the GUI, joystick, the car and all of its functions, are working as we envisioned. There are really a red thread through everything, and we are handling the complexity really well. This is fun and inspiring for us, to say the least!

We implemented a safety function to the joystick mode this week. This function is implemented on the microbit, to reduce the risk of crashing when driving with the joystick. It is not easy to see and operate the car just using the camera, because the difficult to percive depth, and there is a bit of latency on the video feed. Its implemented on the microbit, wich has code for the motors and ultrasonic sensor. This seperates it from the other parts of the system, and reduces response time. This makes it trustable, and highly useful. This is an elegant way to protect the car and its components from damage.

Collage:

Videos:

August:

Well, its that time of the project. Where your hair falls off, and whats left turns gray. Its time for integration hell raised to the power of 2000 factorial. I co-operated with Sander on this one. We learned a lot. Primarily that things (still) takes time, especially when merging. Im also glad that im not an electro-engineering student. The cable work was batshit crazy, but we managed it fine.

Iteration 1 this week on the frame:

Jamal Safi once said, and i quote: “You have to iterate! Dont fall in love with your first idea!”. I now understand what he meant. It turned out that he wasn’t a crazy systems engineer after all. He was right all along. And iterate we did. We needed to get the placeholders to the frame for the camera and the ultrasonic sensors. Our first drawing (earlier week) witnesses our initial plan, to 3d print mounts for the ultrasonic sensors and the zero-cam. We later thought that it would be a great idea to just drill holes for them in the front of the frame. We measured all the sensors and the camera, and sketched a new design:

Iteration 2 on the frame this week

List of everything done this week collaberating with Sander:

• Tried to drill holes for the sensors and camera in the front of the frame of the car. This didnt go well
• Draw new frame design in drawio and sent it to Richard
• Richard fixed the frame design
• Had to do an iteration, one hole was too small in the front. Draw new, and sent to Richard
• The frame was glued together upside down, so we took it apart, and glued it all back together with the correct front
• Taped all the cables for the arduino and the sensors
• Mounted everything to the car, sensors, camera, lidar, rpi4, arduino, motors, wheels, microbit driver board, microbit.
• Merged every ros2 node and the topics. Making it all work on one RPi4.
• Making sure that everything worked
• Spent a lot of time to get all the ros2 nodes to work. This was mainly because of different libraries and packages that one RPi4 had, but the other RPi4 did not.

Summary of Integration Hell

Look at the pictures above, and blogposts from earlier weeks. We have integrated and merged everything. This took time! My god.

We are closing in on our goal!

Next week:

• Iterate on GUI
• Iterate on cable work in car
• Getting new front with placeholders for ledpins

It is imperative to optimize time usage the future weeks.

Sander

This week was spent on integration and iterating on the rover design. I cooperated with August a lot this week as seen in the pictures over. We created a new and better drawing for Richard, where we added more holes, including pre-cut holes for the wheels, this was to reduce the friction on the wheels which seemed to help.

Done this week with August

• Created a better drawing for Richard
• Redid the front panel
• Took the car apart and glued it together
• Mounted everything to the car
• Merged every ros2 packages to 2 workspaces, one for the Pi and one for the PC with the GUI and joystick
• Spent some time merge conflicts specially regarding the camera and joystick, since they was configured on separate Pies. We also had to reconfigure some wifi settings on the pi to get them to work. The problem was that one of the pies cant connect to a screen
• This was finalized in a test driving with all the sensors and camera displaying in the GUI. Everything that we imagined from the start was now realized.

Next week:

• Continue making the car autonomous
• Work on mapping
• Iterate on the front panel to make room for the led pins