Date: 31.10.2019 – 07.11.2019
While waiting for the Raspberry Pi 4 to arrive, we started setting up a Raspberry Pi 3B+ board that we had laying around from before. We flashed it with an operating system and installed all dependencies for the system. We hit a snag when we found out that there was no Raspberry Pi camera support for the AARCH64 architecture, which means we had to wipe our setup and install an ARMv7 operating system instead, which AARCH64 is backwards compatible with. This entire process took a long time and after all this time we concluded that it should work, but we ran out of time to confirm. The plan is to set up the Raspberry Pi 3B+ fully with ROS, TensorFlow and everything, and then clone this operating system onto the new Raspberry Pi 4 once it arrives. This way we can still do some work and get things done without waiting for the new single-board computer to arrive first.
Last week we mentioned that the GPIO pins on the Raspberry Pi can provide up to 1.5A at 5V, which is significantly more than what can be achieved on the Jetson Nano. This allows us to connect our landing gear motors directly to the Pi. We have discussed this and we might keep the Arduino board after all, considering the fact that we already put good effort into making it work using an Arduino and connecting that to the rest of the system. We are currently using an Arduino Nano to control the landing gear motors, and the Arduino does not weight much, so removing it will not affect the drone weight by a large scale. We might gain some benefits such as less cables and no extra power supply, but this is something we will discuss more this week to ensure that we make the right decision.
We have also started to put the drone together. This started with a few test-fittings of a few different parts. Engine-mounts, mounts for the servo and “legs”, mount arms to the body, etc. We also managed to test out the carbon fibre arms we will use. They seem to be of good quality and will handle the forces it has to. We also tested out a small jig we made to drill holes in the “legs” for the landing gear. This worked out as it was supposed to. However, after testing it together with the servo-engines for the landing mechanism, we chose to go for another solution to make the drone more stable when landing and taking off.
Another problem we encountered was lack of hardware. We had ordered screws, spacers and nuts, but since we had a few design-changes, we had to order both other types and more of the ones we had. This isn’t the biggest problem but it delays the process of mounting everything together and testing things.
Since we are making a drone it is very critical that we make the drone as light as possible. After recommendations from Kåre, in terms of making the drone as stable as possible, we wanted to make sure the weight was below 50% of what the maximum thrust of the engines are. Even though we had already made a sheet of technical requirements including weight, we wanted to make sure that the parts we have received the last few weeks was not far heavier than our estimates. To make sure we were as accurate as possible, we wanted to weigh everything ourselves, to make sure the drone would take off. The fact that we chose to weigh every part beforehand came in handy since we are going to be machining out the carbon fiber plates on friday. This made us able to see if we potentially could save more weight on the main body of the drone.
We stumbled upon a few problems with the controller and making it communicate with the drone last week, so we tried again a few times this week. This time we think it works and the drone is ready for arming. Arming is a special combination of the TX-sticks on the controller that must be done to be able to fly the drone. This is for safety reasons only. To be able to go any further on this, we must mount everything together and we need to ensure that the motors are tight before we attempt to arm them. This is a process that we will continue next week, as we are still waiting for some parts to arrive.