During this week I continued to work on my final paper which is about 17 pages long, and prepared for my poster presentation for the symposium event.
During this week I continue with my investigation, this time on redundancy.
I found a publication on the basic understanding on triple modular redundancy for FPGA -based computer systems.
The basic idea is that TMR is when you have three processors working simultaneously, and their outputs have voters. The voters select the most popular output. TM is implemented relative to the architecture structure of the FPGA. The publication explains that the biggest area of the architecture such as the registers for example will require a very strong TMR since it is more vulnerable to radiation effects. This is to avoid overhead and also to avoid to make the system prone to radiation effects.
During this week I continued to work on my final paper, and finished my poster presentation.
During this week I clarified the topics of my investigation that were unclear.
I investigated about the effects of single event burnout (SEB) and single event gate ruptures (SEGR).
These effects mostly affect power MOSFETS. The main idea is when a high energy particle strikes a power MOSFET it alters the electric field at the depletion regions or at the insulator layer. If it accumulated in the depletion regions it creates a SEB, whereas if it accumulates at the insulator gate it creates a SEGR.
I also investigated about SEL.
A single event latch-up SEL. Affects mostly complementary MOS (CMOS). The main idea is that the energy from a high energy particle activates the parasitic behavior of the CMOS which implies that it starts to behave as a Bipolar junction transistor (BJT). Relative to the energy of the particle the current starts to increment leading to a short circuit described as a latch-up.
I also investigated about error detection and correction codes EDAC. These techniques exist on both hardware and software. These are implemented relative to the mission to avoid overhead in computation.
During this week I started to work on the literature review, and on Friday I worked on the draft for the poster.
As I was writing the poster with my own words I realized that the dynamics of the space environment is a subject I feel very comfortable with, but as I had to describe the physics behind all of the different types of SEEs I realized I need to spend more time understanding them in order to present them during the symposium.
I also need to understand how the mechanisms of the soft mitigation techniques deeply work since I cannot explain them right now with my own words.
Two very important concepts I learned during the week were:
- Linear energy transfer LET: Linear transfer is one measurements to account for when measuring SEEs. LETs is the energy a high energy particle or heavy ion dissipates around its environment as it penetrates a semiconductor. As is penetrating this semiconductor it leaves trail of electron holes pairs in the insulator material, many times extending up to the n or p substrate. As this particles go through process of recombination a drift current is created. The density loss is proportional to the density of the charge trail.
- Learned how to read solar particle events from the JPL-91 model:
The solar particle events published by the JPL-91 provides data of 3 solar cycles composed of 11 years each. The graphs provided are useful for solar particle events predictions within a one astronomical unit. The models are graphs of differential solar fluence particles/cm^2/s/sr as a function of the kinetic energy of the solar protons in MeV.
As I continue this work during the Fall semester in my university I realized I need to design a study plan to understand how to account for the total ionizing dose TID accumulated in the microprocessor while the benchmarks are running under the radiation environment of Co-60.
I need to find literature that explains where in memory the benchmarks I have indicate that a problem has been populated.
I also need to design a test plan, that is in accordance with “as low as reasonable achievable” ALARA
During the week I met with Dr. Bradley and we decided he will allow me to decide what type of radiation mitigation techniques I want to use for this investigation. I no longer have to relate to a mitigation technique with his expertise on artificial intelligence and control systems.
What do we have so far?
We want CubeSats to be successful in space despite the fact that they are very sensible to the space environment.
So far there has been hundreds of missions with CubeSats for short missions specially in LEO.
Recently scentists want to take these space systems for interplanetary missions in order to allow universities
and small companies study deep space and dynamic orbits.
NASA successfully completed a missions of two 6U CubeSats named Marco accompanying the INSight Rover.
The publication of this mission specified that the radiation technique used was (TMR), if the voters are encountered with an error
the system resets in less than a second.
Other historycally used radiation mitigation techniques for CubeSats:
Spare tiles, ready to replace tiles affected by a SEU
Other mitigation techniques that could be used:
nanomaterials, but are they really worth it for these type of missions?
Maybe it would be worth it if you would like to use a CubeSat for a longer mission on both environments
LEO or deep Space, or if you would like to explore a more dynamic orbit such as a sun synchronious, or High elliptic orbit.
Items to take into consideration:
understand the dynamics of different known orbits that a cubesat could be deployed at, including the hyperbolic orbit
Where to go from here?
Most of the publications I found explore the mitigation techniques against non-destructive SEE,
but I wonder if there is anything established to mitigate destructive SEE such as SEL, SEB, and SEGR
which occur at the hardware level.
Today I am also going to read this publication:
Taking SmallSats to the Next Level-Sensible radiation requirements and qualification that wont break the bank.
Lets see what type of ideas this could bring, I will also investigate about mitigation techniques at the hardware
level since when these occur the entire system most likely is terminated.
I found an article named: Fraunhofer Satellite radiation sensing systems.
I am really interested on the techniques used for this publication since any type of SEE is initiated at the hardware level.
During week 5 I finished reading two of the publications I selected. One fully explained the radiation tolerance of commercial off the shelf components currently used in CubeSats. It explains how certain components such as a microprocessor cannot be used once its radiation threshold in met, even if you allow the component to anneal for 24 hours in room temperature. while other components are capable to recover after annealing such as the translators to step up or down the voltage required for the microprocessors, and the non-volatile memory EEPROM. All components were tested under a radiation environment while simultaneously running benchmarks to find the single event effects faults.
The second publication from Montana State University MSU, explains the implementation of a satellite’s on board computer using an FPGA divided in nine tiles. The FPGA uses triple modular redundancy, memory scrubbing, and uses a radiation sensor to identify single event effects to replace the affected file faster vs waiting for the memory scrubber to find the discrepancy.
At this point I realized that the rest of the publications talked about the same subject using the same radiation mitigation techniques in different manners.
I decided to keep looking for publications that could provide insights on how to probably use a satellite swarm, a control system capability, or artificial intelligence to aid the radiation mitigation techniques efforts.
I found a publication were it explains the results of the fault tolerance reliability of a partially implemented state vector machine on hardware using an FPGA.
At the end of this week I realized I did not know how to continue doing this search of relating my knowledge on radiation protection with Dr. Bradley’s expertise to make a collaborative project.
During week 4 I spoke with Dr. Bradley on Tuesday 25 June. We decided not to continue on the potential project to use a small drone to execute a search algorithm to find the shortest path out of a radiation zone.
We decided for me to go back to my original plan of investigating current radiation mitigation techniques for CubeSats in low earth orbit and deep space. We were hoping to implement an additional mitigation technique that could utilized Dr.Bradley’s expertise on control systems.
I continued reading multiple publications on current implemented mitigation techniques for CubeSats, and experiment results of post-radiation reliability of certain components such as microprocessors, FPGAs, and non-volatile memory hardware performing operations with and without mitigation techniques under a radiation environment.
I also kept looking for publications of potential radiation mitigation techniques using a satellite swarm, or by using the capabilities of the control system of the CubeSat in a manner to delay the total ionizing dose threshold of the components, and reduce the rate at which potential single event upsets occur.
During week three I read the introduction chapter of the search algorithm where I learned about: breadth first search, and depth first search. I tried reading some of the source code available online but I was not fully understanding the intuition for the source code since I had some gaps about data structures, and how to program in python.
I read about stacks and queues since these are the data structures used for BFS and DFS. I found a book from the University’s library with composed of small videos to learn how to program in python:
I should be done with videos by Sunday.
On Monday I will review again the idea behind the search algorithms from the lectures:
in order to understand the BFS and DFS completely.
During the week as I kept doing investigation on how to complete a mission to deep space with a small satellite, I realized that I really at this point have an specific problem to solve in term of radiation with a cubesat.
I read a publication from Dr. Pellish from NASA in regards to future single event effects and total ionizing dose for autonomous vehicles in space and how this radiation could affect this technology. The mitigation techniques are the same as for a space system (redundancy, and material protection). Since I cannot do materials research at this time, the publication at least inspired me to do my project on a radiation sensing drone.
On Friday I began my investigation on how to implement such drone.
During week 1 I planned to work during the summer on a CubeSat capable of mitigating the radiation of Space.
I was very attracted by the fact that only a few satellites have traveled through deep space. That being said I wanted to plan a mission for the CubeSat in deep Space and investigate how to protect the components of the vehicle to complete its short mission.
The potential mitigation techniques could be materials, redundancy, and escaping Earth’s magnetosphere on time.