Contributions

Architected the backend IoT telemetry pipeline and physical network infrastructure to ingest and log 10Hz flight data with zero-touch deployment capabilities.

Result

The pipeline successfully achieves <50ms latency from hardware to database. Running docker compose up instantly provisions the server, auto-configuring TimescaleDB schemas and EMQX SQL routing rules seamlessly.

Engineered real-time digital signal processing queries and comprehensive mission control Grafana dashboards for live flight monitoring.

Result

Grafana renders live maps and 10Hz dynamic graphs seamlessly. Mach and drift calculations are proven mathematically accurate and visually stable against a custom Python telemetry simulator I wrote to inject mock flight data and stress-test the MQTT-to-database pipeline.

Designed a custom mixed-signal PCB capable of distributing 44V/48V high-current motor power alongside sensitive 3.3V/5V RF and logic circuits without interference.

Result

Power routing, fusing strategy, and ground plane separation pass initial schematic DRC and layout clearance checks. The star ground topology successfully isolates the high-power domain from the logic domain in the physical layout, theoretically eliminating ground loops.

Implemented a dual-microcontroller hardware architecture utilizing two Raspberry Pi Picos to decouple hard-real-time motor kinematics from asynchronous RF telemetry processing.

Result

Schematic logic is fully defined, ensuring safe voltage translation to the motor drivers and correct pin mapping for all hardware SPI/UART peripherals. The architecture theoretically eliminates blocking delays between communications and kinematics.

Interfaces

FromToTypeDescription
Radio PicoMotor PicoelectricalHardwired UART inter-board link passing decoded GPS targets and tracking vectors to decouple RF processing from motor kinematics.
Motor PicoExternal Stepper Driverselectrical3.3V PIO logic signals (Step, Direction, Enable) shifted to clean 5V for robust motor driver triggering.
Radio PicoGround StationdataGround station WiFi connection enables streaming telemtry data directly into ground station's EMQX MQTT message broker.
Fill StationGround StationdataHigh-bandwidth wireless bridge carrying physical pad sensor states and zero-latency go2rtc camera feeds into TimescaleDB and Grafana.

Documents

In-flight Dashboard
IMGIn-flight Dashboard

Demonstration with mock data

PCB in 3D View
IMGPCB in 3D View
PCB Layout in Altium
IMGPCB Layout in Altium
Schematic Sheet 4 - Radio Pico and Peripheral Connections
IMGSchematic Sheet 4 - Radio Pico and Peripheral Connections
Schematic Sheet 3 - Motor Pico and Peripheral Connections
IMGSchematic Sheet 3 - Motor Pico and Peripheral Connections
Schematic Sheet 2 - High to Low Voltage Conversion
IMGSchematic Sheet 2 - High to Low Voltage Conversion
Schematic Sheet 1 - 48V Power Entry and Distribution
IMGSchematic Sheet 1 - 48V Power Entry and Distribution
Our ground station setup in the field
IMGOur ground station setup in the field
RATS System Architecture Diagram
DGMRATS System Architecture Diagram

Limitations

  • PCB design, layout, and assembly are complete, but comprehensive bench testing is pending.
  • True Hardware-in-the-Loop testing of ground station has not been conducted.
  • Real-world RF multipath fading and environmental packet loss metrics at high altitudes have not yet been characterized.

Lessons & Next Steps

  • Investing the time upfront to build the mission control backend entirely as Infrastructure as Code (Docker Compose) saved countless hours during field deployment testing.
  • If I were to do this again, I'd try to stick to a more uniform component list for my hardware design - having over 80 unique components on a single board is not easy to manage, both in terms of handling inventory and assembly.

Built something worth showing?

Document your project →