Technical
- Redundancy requirements
- Actuator requirements generally met by adding a 4th reaction wheel. Providing redundancy in the reaction wheels ensures the fine pointing accuracy requirements of the mission can still be met. The coarser pointing accuracy of the magnetorquers would lead to premature mission termination if a reaction wheel were to fail.
- Three power options are available: Cold, Warm, and Hot backup configurations. Cold configurations keep the redundant actuator off and turn it on when it is needed, warm configurations keep it powered but it is not in use until another component fails, hot backup has all four components powered and contributing to the actuation.
- Hot backup is ideal, as it gives increased control authority. Related drawbacks include increased power consumption.
- Two common configurations are available when a fourth reaction wheel is introduced
- NASA standard configuration → one RW pointing along each axis, fourth RW at angle to all 3 axes
- Pyramid configuration → All 4 RW at an angle from nadir, ideal for Earth-oriented missions, usually require a hot backup power configuration
- Sensor requirements are more ambiguous, essentially add more sensors to take over when sensors fail. There is an option to create a redundancy through an algorithm, though this concept is currently not too well understood by the ADCS team. There is currently no plan to pursue the algorithm redundancy, though it may be a viable consideration for future designs.
- Algorithm creation
- Translation from C++ to C is a lot more complex than it seems. Leave a significant amount of time and resources available to properly accomplish this.
- All algorithms should be added directly to the UW Orbital main firmware repository so that compilation can be tested for everything all at once.
- We should be aiming to use MATLAB/Simulink to test out all mathematical control laws and algorithms prior to touching any C code. This allows us easy integration with out simulation environment and can rapidly test things out. Plus MATLAB support testing out the mathematics right out of the box instead of needing to tediously code it in C first.
- We can use MATLAB live scripts to include the typeset mathematics powering the algorithms we are writing directly in a MATLAB environment. This allows us to bundle our test code alongside the mathematical definition of what we are hoping to achieve
- Simulations
- Aimed to use STK SOLIS to verify architecture. However, SOLIS is found to be developed by a third party and cannot be included in a sponsorship package.
- STK can also be used as a visualization tool to visualize MATLAB output. This effort is ongoing and is yet to be proven. Through this method, STK can be used to visualize and verify algorithms developed.
- All simulations, at least with respect to validating control laws, need to be tested in Simulink. This allows us to create mathematical models of how the dynamics of the CubeSat should behave and how allows us to include various environmental models into the simulation.
- We should be typesetting mathematics or somehow including the mathematical knowledge within the simulink environment so simulink blocks are more obvious as to what they are intending to achieve
- Actuator sizing and environmental disturbances
General Management
- Leave 1-2 weeks buffer on tasks to account for unexpected circumstances (i.e. projects, sickness, etc.)
- Move mid-week updates and general discussions to specific channels on UW Orbital discord. Old method of PM’ing members for updates kept other members/leads in the dark with regard to task progress.
- Problem moving to UW Orbital discord → how to hold multiple parallel conversations at the same time?
- Create more channels? → discord might get too crowded, need to find a compromise
- Need to make better use of task board on Notion → use spring cleaning to cut down on amount of tasks (currently overwhelming). Make sure to keep tasks up to date (completed vs. incomplete).
- One problem → how to deal with subtasks in Notion?
- We need to potentially write up a series of tutorials on controls engineering and how to mathematically model and simulate various hardware and sensors → this is an integral part of the simulation we want to run → need this knowledge
- Maybe we need to run workshops where we can give out some training for useful math and programs to ADCS members, this might help form a common knowledge base within the sub team earlier in the term