Blog 9 - Final Capstone Blog!

The last three weeks have been very exciting for our team because testing has begun with integration of our electronic components with frame components from the Mechanical Engineering team. For initial fitting and testing, the frame components were 3d printed using PLA (PolyLactic Acid) filament. In the first fitting, we realized the motor supports needed a center indentation for the shaft of the motor to rotate freely, so that was added to the CAD design to be implemented in carbon fiber for the final prototype. The 3d printed drone's initial assembly is pictured in Fig. 1.

The electronics box was then finalized in CAD and printed for fitting. Initially, the Raspberry Pi did not fully fit in the box with the motor ESC and flight controller, so the first flight tests were performed without the camera to get a feel for how the drone will handle. The redesigned electronics box will hold the Raspberry Pi, camera, and spotlight and is expected to be finished and printed this week.

As the drone is designed for efficiency over power, the weight of the drone is an important consideration. Upon further review of the drone's structural stability, the foam blade guards were removed to reduce weight improving the flight time and drone maneuverability. The blades will be sufficiently protected by the PLA filament guards shown in Fig. 1.

The flight controller settings from Velocidrone were implemented on the drone. The controls from the Betaflight PID control were converted for Inav, as the PID controls are slightly different for each software. Shown in Fig. 2 are some changes from Betaflight that were done for Inav. More flight testing will be done with the drone to continue testing stability and maneuverability.

This image is available upon request by only CAPSTONE instructors. 

Fig. 2 - Changes to the PID Controller from Betaflight to Inav.

Since the previous blog, there has been a change in both the operating system and priorities for video transmission using the Raspberry Pi. Since the drone was in the process of assembly, efforts have been redirected towards allowing the barebones functionalities. If time allows, further features will be implemented.

The Raspberry Pi's operating system has been switched from its full desktop OS to its Lite Os, which consisted of simply a command prompt. This allowed for more computation resources available for the video software as desktop rendering and other background processes are now omitted. The result is new, barebones command window that the user can use to run commands and render the video stream on top of.

As required by the deliverables, the drone must be able to begin video transmission once it receives power. This functionality requires modifications to the drone's software. Tasks must be created in its operating system to run the transmission service. This problem was solved using a series of solutions. First, the user authentication must be disabled, as the drone will refuse to run code until access has been granted. This was solved by going into the OS's configuration menu and disable auto-login. Next, the drone is required to run a specific process (libcamera) in order to begin video transmission. For this solution, a bash file containing a series of console commands is executed after login via bashrc. Overall, with these two functions in place, the raspberry pi will start up its video transmission immediately after it's powered on and goes through a quick setup. The video output with analytic information is shown in Fig. 3.

Fig. 3 - Video output from Raspberry Pi with the night vision camera and optimized software.

While the tasks may seem small and simple, this is a result of many attempts of research and trial and errors, as this exact process hasn't been documented at all online. This is one of the shortcomings of the team's software work, as the original estimation of the software development was wrong. However, current assessments indicate that the software should be ready by the next demonstration deadline.

We were able to have the carbon fiber plates manufactured and cleaned in preparation of the final assembly and testing. The parts are shown in Fig. 4, and a big thank you to the HCC Innovation Studio shown in Fig. 5, for their tools, space, and expertise! They helped us get the last piece of our final project off of the ground, and we're very thankful for that. As mentioned earlier, we were able to fly the PLA-printed prototype for initial testing, and we look forward to continuing testing with the finalized electronics box and carbon fiber frame in the coming days and weeks, even after post graduation.

-Matt Andersen, Graduating Class of 2022.

Founder's Note - 

As a project, this has been an ongoing effort for a very, very long time. And after this capstone, I think the work we have done is almost ready for the big leagues. I think that we have just reached the point where we can succeed in the real world, and will be ready to succeed by the end of the year. I am looking forward to when that happens.

--Bernard.