GNSS receiver

This GNSS receiver is an ongoing solo project which aims to construct an entirely custom L1 band GPS receiver for use with high-power rocketry. This receiver will allow accurate tracking of launch vehicles traveling at high velocities and altitudes, at a substantially lower cost than commercial solutions. This project has required a large amount of self-instruction, requiring advanced skills in RF design, satellite communications, PCB design, embedded systems design, and FPGA systems.

Pictured below are the first revision test modules of the receiver, an RF front end for signal amplification, filtering, and downconversion, and a fractional-N PLL frequency synthesiser which acts as the systems local oscillator. These systems operate in the microwave frequency range, at approximately 1.6 GHz. This is a long-term project, after testing the above subsystems will be combined into an integrated receiver which will handle the computational aspects of the system.

As with all complex RF systems, multiple hardware revisions are needed to get optimal results. In the above LO design, not enough care was taken when designing the loop filter and output matching circuits, leading to unsatisfactory phase noise performance (the requirements for which are stringent in a GPS reciever). The front end was unable to function due to unmitigated DC offsets building along the signal path to the comparator, and a data acquisition scheme whose bandwidth was too low to properly capture all needed information. This latter problem was due to a lack of accessibility to proper test equipment.

All the above issues have been addressed in updated versions; the RF front-end is shown below. A new LO generator was chosen for improved phase noise performance, and greater care was taken in designing the surrounding hardware. The RF front-end has had a complete overhaul, switching to a balanced IF section and a differential ADC with an FPGA-friendly parallel data interface. These new boards will be tested using an upgraded digitisation pipeline, which should be more representative of how data will be captured in the final reciever.

The primary problem I’m trying to address with this project is the fact that commercial GPS recievers cannot be used on high-altitude rocket flights. High altitude and space-capable GPS trackers are exorbitantly expensive, and unobtainable for student rocketry teams. I few years ago I decided that I could simply design my own system, and make it as inexpensive as possible. The design is made up of several elements:

• RF Front End - this section is responsible for recieving signals from the GPS antenna. Received signals in the L1 band are approximately 20dB below the thermal noise floor, so must be treated carefully. The RF section amplifies, filters, and downconverts the signal to a 20MHz IF. The whole section is designed from readily available, non-specialised components.
• Local Oscillator - this section generates an LO used in downconverting the signal to its IF. It needs to be a highly stable, low phase noise design, as any drifting of the LO will result in potential loss of GPS lock in the tracking loops, and a larger search space for doppler shift determination during initial acquisition. This is of particular concern in rocketry applications; frequency references are often vibration susceptible so care must be taken to ensure immunity from environmental factors.
• Digital Section - the digital section takes the processed IF and extracts the navigation data in order to compute a position solution. This first requires identification of visible satellites and determination of carrier and code phases. This is a computationally intensive process, relying on an FFT-based algorithm which marches through an enourmous search spaces of frequency shifts, code offsets, and SV IDs. Next, visible satellites are tracked using simple control loops. A Costas loop extracts navigation bits from the incoming data, and the other loops adjust the system timing to account for each satellite’s changing doppler shift. The extracted navigation data is parsed and fed into a position algorithm, which determines the receiver’s location using the each satellite’s ephemeris data. The digital section additionally decides which satellites to track and which to ignore (to minimise dilution of precision), and manages data storage and communication.

I am in the process of designing a testing suite for my new boards, which I hope to complete by the end of January 2026. I’m looking to validate that I can recieve signals, and that those signals have sufficient SNR for further downstream processing. Once this is confirmed, I’ll move on to integrating the RF sections with a digital processing system to create a compact integrated system.

Whilst this project is far from completed, I have already learned enourmous amounts from the work I have done. This is my first proper RF project, and has been a wonderful challenge to wrap my head around the advanced concepts involved in satellite positioning and RF electronics. I’m excited to continue the development of the project, and would love to one day fly it on a rocket into space.