126 - APPROXIMATE COMPUTING FOR LOW-POWER GNSS SIGNAL PROCESSING

About the role

Global Navigation Satellite Systems (GNSS) provide accurate positioning, navigation, and timing information for applications ranging from mobile phones to infrastructure and military activities. GNSS receivers require significant computational resources, particularly during signal correlation. This project explores Approximate Computing (AxC) techniques to reduce power consumption while maintaining sufficient accuracy in GNSS signal processing, benefiting energy-efficient platforms like consumer electronics, wearables, IoT devices, and spaceborne receivers.

Responsibilities

• Identify the most relevant AxC techniques for GNSS signal processing.
• Develop a prototype GNSS receiver using AxC and exact computations, supporting both nominal and AxC modes.
• Design and implement:
  • A bit grabber to collect IQ samples from live data and laboratory playback (e.g., using USRP and a GNSS antenna).
  • Test equipment to control and monitor the GNSS receiver and bit grabber.
• Conduct tests to characterize receiver performance in terms of power consumption, accuracy, and correlation degradation.
• Utilize provided data from ESA Navigation Laboratory to replay scenarios in open sky, urban, and jamming environments.

Deliverables

• Technical notes describing tradeoffs and design decisions.
• A functional prototype of a GNSS receiver incorporating AxC (covering acquisition, tracking, and PVT stages). The prototype does not need to operate in real time but must support both nominal and AxC modes.