One of our engineering sub-teams here at Rutgers RPL is the Telemetry team. This group of ten individuals is led by team-lead Aaron Williams and is responsible for the collection and transmission of all data during flight. This includes transmitting mission-critical information such as current altitude and GPS position.
Orientation Algorithm
Currently, the sub-team is in the process of implementing an orientation algorithm that they developed. Essentially, it takes gyroscope and accelerometer data of a sensor, which is located on a circuit board, and computes the orientation of the board relative to the ground while it is inside the rocket. Using the center of mass of the rocket, coupled with the distance to the board, as well as a predictive model of the flight profile for each individual rocket, we can quantify the rotation of the rocket.
Historically for this program, the microcontroller that powers the sensor computes the rotation as well. However, now that we have a raspberry pi (a tiny computer) as our flight computer, we want to offload the computational work onto the new microprocessor. Now in an attempt to visualize the orientation, Telemetry is testing the technology on a computer that runs Python, while piping the output into Processing (Java).
Importance
In terms of two-stage rockets, if you are able to find the orientation of your rocket, then you can compute your pitch and therefore know how far you are from being vertical. Even if you cannot physically see the rocket, you can send four numbers to it and use Java to identify what orientation the rocket is currently at. If the second stage is tilted in an unfavorable way, you can set conditions that tell the motor not to ignite it. This is incredibly important for safety! By “tilted in an unfavorable way” we really mean “tilted in an unsafe way.”
Future Outlook
So far, the Telemetry team has only run static tests of the program. Going forward, the team is working on finding a way to accurately test the algorithm on our rockets. The transition has proved itself difficult because the program partially relies on knowing where gravity is pointing. Accelerometer data cannot be used when a rocket is accelerating or in freefall because an external force is acting on the rocket on the former case, and the accelerometer reads zero acceleration in the latter case as the board of the accelerometer is accelerating at the same rate as the accelerometer itself. So the way they are planning to make the program work under real-rocket conditions would be to make a highly accurate predictive model of all the forces on the rocket, taking into account as many factors as possible, essentially creating a complete mathematical representation of how the flight should go, and then during flight, comparing the real sensor values against their modeled counterparts to weigh the error and say with a degree of certainty what orientation the rocket is at.
Thanks for keeping updated with Rutgers Rocket Propulsion Laboratory!
Until next time,
Per aspera ad astra- Through hardships to the stars
– Maya Ziab, Director of Social Media
– Aaron Williams, Telemetry Lead
Rutgers RPL 