• Created by Unknown User (santora), last modified on 14.08.2018

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 3 Next »

Overview

Visit to UGA SmallSat Research Lab in order to understand the overlaps of two very similar missions (theirs and ours) as well as understanding how they run their lab, program management, student management and transfer of knowledge. How we can collaborate and share ideas.

Lab: Clean room, EDS, FlatSat

Team communication: Google Drive, Slack

Team hierarchy: PM->leads->subsystems & faculty on the side

Minutes

Present: Caleb Adams, Nick Hollis, David Cotten, Deepak Mishra, Mariusz

08:30-09:15

Programmatics & project management

How lab started and what it is made of

  • Started in 2015 with initiative by David and Caleb. Received funding from NASA to create a mission. Proposal sent by Deepak and David to do coastal monitoring from space (hyperspectral imagery) and won the competition between several universities.
  • SSRL is initiated by Dept. of Geography and Dept. of Physics, approved by Dean. The physical lab is in Physics building.
  • Team size fluctuates between 30-50 people (students, faculty, summer interns) throughout the year
  • The SmallSat Research lab is purely created and run by volunteering students
    • There is no research credit given for undergrads now but David wishes to employ this such that people get trained
    • Only paid people are summer interns – which are there throughout the summer to have continuous flow
    • Faculty only work as advisors, not bosses. PM and leads work as “bosses” and run the lab.
  • David teaches a course about CubeSats at UGA
  • UGA does not have Aerospace Engineering department like Georgia Tech – hence SSRL is more interdisciplinary (and potentially more creative).
  • The students background mainly come from Computer Science, Mechanical Eng., Electrical Eng., Advertisement, Arts – Please see Personnel Budget Document.
    • You would be surprised how students from Advertisement, Arts etc. can contribute!
    • They have no biology/remote sensing expert – Deepak Mishra is (faculty)
  • Total cost excluding launch costs for the 2 missions is about $200,000 to this day (mostly HW+facilities+travel). 

Interviews

  • They have 3 round process of interview once a semester.
    • Get perhaps 200 applicants
    • 50 % are immediately cut off – looking for skilled people that are needed for specific tasks that are advertised
    • GPA doesn’t matter that much. Normally varies between GPA 2.0-4.0 for accepted students.
  • Second and last round are more personal interviews
    • 5 % acceptance rate – i.e. they select normally only 8-12 out of 200.
    • First and foremost it is passion that counts most as well as team-working skills and social abilities
    • It helps if they have done some significance in extracurricular activities and developed something – e.g. Arduino projects, computers etc.

How they run the lab

  • Project Manager solely runs the lab – ownership of the projects are with the students, not the faculty
  • Lab students normally work 10 hrs a day (incl. UGA lectures+coursework), but all is voluntary
  • Faculty gives advice and keeps in-check where critically necessary (conflicts, low resources)
    • Confidence given to Caleb and Nick Hollis (PM and chief engineer).
    • For mission design they follow the NASA/ESA standard with some short-cuts though mainly:
      • Requirements
        • Define science requirements (Deepak gave a lot of these)
        • Define mission objectives
        • Define mission success criteria
        • Define mission requirements (functional & non-functional) & constraints
        • System Requirements
  • Reviews (mandatory when funded by NASA and US Air Force)
    • Mission Definition Review
    • PDR
    • CDR
    • They get tricky questions but exteriors always try to simplify things – very helpful feedback
    • No delays to this date
  • Between each formal review they run delta-reviews. These are:
    • Internal reviews
    • Follow a standard template
    • Trains them and prepares them
  • Train STEM students 5 times a year
    • Supported by NASA
    • But mainly SSRL’s passion
    • Well appreciated in the community
    • Teach: how to build Arduino, how to build cubeSat etc.
  • Motto 1: “Have fun”
    • They have couch and video games in the lab (David was against it but it works even better!)
    • Makes people stay longer and being informal (also with faculty – low-level communications is very important)
    • Social activities
  • Motto 2: “Ask a lot of questions”
    • Even "stupid" ones
    • Caleb asks a lot of questionsBe curiousThey use slack for mobile/online communications
      • Faculty is not involved in all channels, they don’t want to – it can limit things.
      • They have their own “fun” channel – much like our “random” channel.

Transfer & keeping knowledge

  • They document very well
  • Have only 1 source of sharing: Google Drive
  • Well structured, Nick Hollis and Caleb constantly tidy it up, structure it and manage what goes where
  • A folder for internal and formal reviews
  • Leaving students write a exit report (status, what has been done and what has to be done) – see template
  • New recruited students are referred to the review folder – so they don’t get overwhelmed plus the exit reports.

09:15-10:10

Nick Hollis presents on overview of the SPOC mission. Deepak Mishra joins meeting. Mariusz presents overview of HYPSO mission. See technical discussions below.

10:00-11:15

Kanna Rajan joins and discussions are about project flow, programmatics & project management

  • See above for outcomes of discussion (most repeated during this chat)
  • Will set up meeting between Evelyn and Caleb (both PMs)

 

11:15-12:00

Technical Discussions

SPOC:

  • They had their CDR with NASA 2 weeks ago (exterior reviewers necessary) – given full GO
  • Coastal monitoring of phytoplankton, HABs, sediments, cyanobacteria (estimate CO2 as well)
  • Camera:
    • Built it themselves with major help from optics experts at NASA Goddard (GSFC)
    • They have 433-866 nm, using only 16 spectral bands
    • Hyperspectral -> works as multispectral (smart imager)
    • Similar to our HSI
    • Use reference of desert in Libya for in-orbit raw calibration
    • They don’t have SNR calculations
  • Use a RGB camera similar to FinderScope (on SeaHawk) to georeferenced/validate the hyperspectral data (in the same housing)
    • Use MicroCam (COTS)
  • Targets:
    • Mainly Georgia coast
    • Deepak has in-situ validation assets
      • Stations
      • UAVs
    • Also interested in Mediterranean
    • They will be in ISS orbit, so cannot reach Norway
    • We should coordinate what targets to look at together
  • Flight Software: KubeOS, MajorTom (https://www.kubos.com)
    • Does not interface with all subsystems (e.g. NanoAvionics or those that run on CSP client)
    • Interfaces with Pumpkin and some ClydeSpace COTS as well as GomSpace EPS
  • They do only Nadir-viewing imaging GSD=130 m. They can tilt using ADCS to look at different angles of same target and compare Nadir vs. tilted imaging.
  • Ground system:
    • Downlink: They downlink payload data on S-band and TM on UHF (I told them why not all in one packet thru S-band?)
    • We can use their UHF Ground station if we would like to (S-band TBD).
    • 9600 baud for uplink
    • 2 Megabaud for downlink
    • QPSK à 4 bits/assemble
    • Don’t have S-band yet
    • $40,000 for UHF hardware and equipment only (covered by internal funding)
    • (I don’t remember/forgot to ask them what Ground Stations they are using, I think they want to use NASA Goddard)
    • Comms. through AX25 protocol
    • Use GnuRadio currently to train with
    • Will use amateur bands in UHF
    • Frequency allocation took shorter than expected - since they have good contacts with NASA and US Air Force (smile)
    • HDR vs. SDR
      • Hardware-fixed bands for space
      • Both hardware-fixed and SDR capability on ground

MOCI:

  • Will perform target detection in coastal areas – high-performance computer
  • The spectral anomalies in the water that are detected are downlinked to ground
  • Two modes: 1) low spatial resolution and high SNR, 2) high spatial resolution and low SNR
  • Currently doing GPU validation for use in space – they have a graduate student working on this
  • Use super-resolution algorithms
    • Artifically noise a set of low-resolution images -> denoise -> output 1 high-resolution image
    • They can help us
  • Use machine learning (neural net) to train spectral signatures to look for – the trained set is used on the payload processor (made pre-flight and fixed onboard). Match/no-match of signatures? Have not thought about updating the trained set yet. Much like remote sensing data are trained in neural network.
  • This is how MOCI ties with SPOC in two missions
    • We (NTNU) want to do this on one satellite and one mission

Discussion with Dr. Thomas L Mote who joined shortly:

  • He is interested in monitoring of coastal regions in Greenland, talks about how arctic upper water column changes with climate getting warmer and SST rising. Talks about field campaigns in Svalbard (Geir Johnsen et al.)
  • Run-offs of mass of water
  • How can we optically observe sediments, subglacial matter in the coastal water – during spring and summer
  • Greenland only has ca. 20 weather stations
  • He would be interested in data from here – currently poorly observed. Even better if we have our in-situ measurements also done here (UAV, ASVs).

12:00-1:00 PM

Lunch with David Cotton and Deepak Mishra

 

1:00-2:30 PM

Tour at the SmallSat Research Lab

Lab configuration:

  • Couch, “fun area”, posters
  • 10-12 PCs
  • Clean Room (Class 1000, ISO6 standard)
    • They received this for free from Physics department
    • Some minor modifications
    • They do everything related to the cubesat here (except for FlatSat), incl. the optics!
  • EDS area
    • Tables for electronics
    • FlatSat configuration (all components except for ADCS and solar panels)
    • They follow Lean procedure for tools and equipment in the drawers, shelves
      • Label everything
      • Stickers on everything
      • Put back immediately after done
  • Ground station area (PC, servers, radios)
  • Brainstorm/discussion area with 2 white boards
  • They have a T-VAC chamber in the works (remodified pressure chamber)
  • They DON’T have Helmholtz coils and ADCS testbeds – need to do this

Pictures

 

Summary

  • Project is almost fully student-led and student-driven, voluntarily, passionate about what they do. Student ownership emphasized.
  • Faculty act as advisors. They interfere when critical (conflicts). Informal communications - sense of a team. 
  • Low acceptance rate. 200 applicants per semester. 8-12/200 accepted through 3 round-process interview.
  • Students (incl. interns) work throughout the year (no extensive/formal holidays) - always somebody in the lab
  • They will support us with both technical documentation and how to run the lab
    • They will provide Clean Room and ESD documents 
    • Mechanical + thermal analysis
    • Concept of operations document
  • We should coordinate overlapping mission requirements
    • Targets
    • Use of ground stations
    • Data processing
  • SPOC+MOCI essentially works in synergy as our HYPSO mission will
  • Help them with ADCS testing (HW)
    • Connect them with G-NAT at NASA Ames
    • Give them material from NUTS
  • Share ideas on:
    • Data processing chain
    • Super-resolution
    • ADCS
  • HYPSO and SPOC+MOCI should act as inter-calibration platforms
  • Evelyn should set up a meeting with Caleb regarding programmatics and project management

Collaboration domains

  • Science objectives, what to observe
  • Greenland aspect
  • Monitor Georgia coastline, give data products to them
  • Connect them to our science board (Deepak knows many of these)
  • Ground Station

Follow-ups:

Caleb:

  • Template/example for exit reports
  • Presentation slides (overview of mission, CONOPS etc.)
  • Documentation on thermal+mechanical analysis.
  • Personnel management report
  • Clean Room + EDS documentation
  • Timeline chart for each mission?
  • High-level description for FlatSat arrangement?
  • Meet with Evelyn over Skype

Mariusz sends:

  • Slides: Overview of HYPSO & mission design
  • ADCS testing procedures + Overview of GNAT test lab and connect them with GNAT, though they are limited to only using sensors and not actuators at the moment.
  • Send alternative to atmospheric correction algorithms (Rick Stumpf's methods).
  • Our list of our science board (please let me know if you want to get in touch with any of them)
  • Concept paper when mature
  • Work on super-resolution and data processing when we get there
  • Check out KubOS flight software: https://www.kubos.com

 

Description of SPOC: 

“The SPectral Ocean Color (SPOC) satellite mission, was funded through NASA’s second iteration of the Undergraduate Student Instrument Project (USIP) in mid 2016.The SPOC's mission is to acquire moderate resolution imagery across a wide range of spectral bands to monitor coastal ecosystems and ocean color. SPOC will acquire image data between 433 and 866 nm to monitor 1) coastal wetlands status, 2) estuarine water quality including wetland biophysical characteristics and phytoplankton dynamics, and 3) near-coastal ocean productivity. SPOC shall use multispectral remote sensing techniques to quantify vegetation health, primary productivity, ocean productivity, suspended sediments, and organic matter in coastal regions. The uniqueness of SPOC lies in its payload, a 16 band adjustable multi spectral imager, called SPOC eye. The payload structure for SPOC was designed by Cloudland Instruments and the UGA SSRL SPOC team is actively building and refining the optical structure. Launch in Q4 2019.”

 

Description of MOCI:

“The Multi-view Onboard Computational Imager (MOCI) mission will acquire imagery of the Earth’s surface from (Low Earth Orbit) LEO and perform near real time Structure from Motion (SfM) at a landscape scale using custom algorithms and off the shelf, high performance computational units. The MOCI mission will also identify and map coastal phenomena such as sediment plumes and algal blooms while training students in STEM related fields. Efficient data compression, feature detection, feature matching, and SfM processing techniques of space-based imagery will be performed on board the spacecraft as a proof-of-concept of high performance, on board processing capabilities. 3D models produced by the MOCI satellite will take the form of Digital Surface Models (DSM) as their end product for quick data downlink. In early 2018 the student team in the UGA SSRL was selected as one of only two winners of the US Air Force Research Laboratory University Nanosatellite Program (Competition NS-9) Flight Selection Review for our MOCI (Multi-view Onboard Computational Imager) spacecraft. The team is now actively building MOCI and is in Phase B of the process. Expected delivery to launch in Q2 2020.”

 

  • No labels