Our final drone design is an ongoing process. As of now, we recommend the Octocopter to be retrofitted with the drop mechanism and refitted with TI electronics. The firmware development required to make the MSP432 a robust flight controller will likely take several semesters. As of now, we have programmed the drone to receive commands via two input button pushes. These commands will raise and lower the throttle of the motors by 25% from 0 to 100 back to 0. Through a separate program, the drone will receive all of the sensory data required to run as an inertial navigation system. When these two programs were merged together, we ran into a few clocking speed misalignments (which led to a dangerous lack of motor controls) and were unable to overcome this obstacle before the final presentation. The future teams will need to consult TI’s E2E forum to move forward. Following that, they will need to receive inputs from a controller to throttle the motors. Once every input is available, the long process of recreating a flight controller begins.
Units in mm.
Throughout the year, we ran multiple tests and had plans for many more before the crash. Our tests and resulting conclusions can be found here.
Regarding development on the MSP432, we recommend that future teams continue utilizing Code Composer Studio (CCS) to develop code for flight control. It has proven to be the most stable and reliable IDE over IAR and Energia. Regardless, development is still a chore, and few resources are available for learning. Documentation is written at the reading level of someone who has extensive prior microcontroller development experience. Using TI’s E2E forums is the best way to get reliable answers to questions.
Potential Future Improvements
The “Broken Arrow” test platform serves as a low-cost, short-turnaround solution for testing new code on the MSP432. Its flight capabilities, however, are not up to par with build-it-yourself drone kits on the market. Thus, we may seek to further improve this test platform by reinforcing weak points in the construction, pursuing lighter materials, and shifting weight around in the design. With enough funding, using a ready-made kit with lots of extra parts instead for testing may yield better results.
The original scope of the project was too large for one team to complete in the time span allotted. Some of the components can be split down into further projects with the following independent teams. Teams 3 and 4 could work at the same time especially if Broken Arrow is still configured to fly.
- The next logical project that follows up on our work is to integrate sensor data and radio control into the MSP432-controlled Broken Arrow platform.
- Once that is complete, and the Broken Arrow drone can demonstrate stable flight and hover without payload, another team will have to scale the programming up to control the 8-motor full-size octocopter carrying the desired payload and implement the pizza drop mechanism.
- Next, a team can focus on computer vision and training a camera to recognize unique fiducial targets for accurate pizza drop once the drone arrives at a GPS location.
- Another team will need to focus on long distance telemetry and control systems.
- Another team will also need to integrate autopilot functionality into the MSP432 flight control system. This team will most likely take advantage of existing open-source autopilot ground control station software for accurate and extensive autopilot options.
- Lastly, a team needs to integrate all systems into the full size Pegasus octacopter factoring in the battery life, weight of all components, and current legal specifications for drones. This team will need to tackle component stress testing, environmental testing, and refining the production and safety of the drone. All systems supporting the autopilot need to work without fail.
It is our recommendation as well that all future teams working on code have a mentor who has extensive experience working with the MSP432 MCU and its Launchpad kits. This would make it easier on future students working on the project to get started and find the resources.
Over the course of this year we made significant progress on a MSP-432 controled pizza delivery drone. We learned that it is better to test out the flight control on a smaller drone to prevent costly repairs on a full size drone. We determined that a pull-pin drop mechanism is the best way to transport the pizza from the drone to the fiducial.
For an exact list of either drones’ parts and their specs, please follow this link.