Tufts CubeSat Team

Building Tufts' first ever satellite.

SPACE RACCOON is a 2U CubeSat mission using onboard computer vision and machine learning to detect, classify, and assess collision risk in Low Earth Orbit.

CubeSat form factor
2U
CubeSat form factor
Low Earth Orbit
LEO
Low Earth Orbit
Mission duration
14 mo
Mission duration
Target launch
2029
Target launch
SPACE RACCOON CubeSat CAD model with exploded internal stack

The mission

From launch pad to open data.

01 / 07 · Target 2029

Launch

After two to three years of design and testing on campus, SPACE RACCOON gets to orbit!

02 / 07 · Low Earth Orbit · 14 months

Reach orbit

Deployed into Low Earth Orbit for a 14-month mission, the satellite runs on body-mounted solar cells.

03 / 07 · Vision payload

Detect

During scheduled imaging windows a modified GoPro captures a frame every 30 seconds, searching for debris against the star field.

04 / 07 · Onboard computer vision

Classify

Image processing isolates each object and a neural network classifies it (fragment, rocket body, defunct satellite) on the spacecraft.

05 / 07 · Collision screening

Assess the risk

Tracks across frames feed a collision-risk assessment that flags the objects that matter most to anything else flying in LEO.

06 / 07 · VHF / UHF

Downlink

Only frames that contain debris are compressed, packetized, and sent over our radio link to a student-built ground station.

07 / 07 · Open source

Open to everyone

Flight software and debris data are all open-source to support safer orbital operations and the student teams that come after us!

FRAGMENT 0.94DEFUNCT SAT 0.88ROCKET BODY 0.81FRAGMENT 0.76COLLISION RISKHIGH · DEFUNCT SAT</> OPEN DATA

Inside SPACE RACCOON

Manufactured on campus.

A 2U CubeSat is just 10 × 10 × 22.7 cm. Inside that volume sits everything the mission needs: structure, power, flight software and vision, and the radio link.

Antenna deckVHF/UHF transceiverPower system & batteryADCS · reaction wheelsFlight computer & FPGAGoPro vision payload2U aluminum chassisSolar panels

Roadmap

26-27 Timeline:

Full timeline →
  1. September

    • Avionics board complete
  2. October

    • Image processing software complete
    • Avionics board tested
  3. November

    • Structure design prototype manufactured
  4. December

    • ADCS board design complete
  5. February

    • Flight software complete
    • Reaction wheel manufacturing complete
  6. March

    • More to come…

Where we and our alumni work

Skills from CubeSat carry into industry. We also collaborate with MIT and UMass Lowell.

  • Medtronic
  • Draper
  • NVIDIA
  • Blue Origin
  • Takeda
  • MITRE
  • Amazon Robotics
EmailWilliam.Goldman@tufts.edu
LocationHalligan Hall, Tufts University
GitHubTufts-CubeSat

Last updated: November 2, 2026