Previous Abstract Return to Session G2a

Session G2a: 40th Anniversary Track: STUDENT SESSION

Lander-Aided Differential Doppler and One-Way Ranging for Lunar Rover Navigation with a Single Relay Satellite
Kaila M. Y. Coimbra and Grace Gao, Stanford University
Location: Windsong 1-2
Date/Time: Wednesday, Sep. 16, 2:58 p.m.

International space agencies and commercial companies are planning a range of lunar missions, from deploying lunar communications and navigation satellites in orbit (Giordano et al., 2023) to returning humans to the lunar surface (Watson-Morgan, 2023). In the near term, many efforts emphasize exploratory robotic missions in the lunar south polar region as precursors to sustained surface operations. These architectures frequently pair a landed element with one or more spacecraft in lunar orbit that provide communications relay and operational support for surface assets. For example, NASA’s Artemis campaign pursues a Human Landing System that transfers to a lunar orbit prior to descent, and commercial providers including Blue Origin have been contracted to deliver crew and cargo landers for lunar surface operations (Watson-Morgan, 2023). This lander-and-orbit configuration motivates leveraging existing mission infrastructure for rover positioning and navigation. For early-stage rover missions, precise absolute localization is needed to support long-range traverses, science targeting, and safe operations in challenging terrain. In this work, we evaluate the localization performance of a rover using three sources of measurements: one-way Doppler shift from the downlink communication signal of a relay satellite, one-way ranging measurements from the lander that deployed the rover, and the rover’s onboard odometry module.
RELATED WORKS AND LIMITATIONS
Many works have investigated vision-based lunar rover navigation, including crater-based map matching and factor graph methods (Daftry et al., 2023; Cauligi et al., 2023; Dai et al., 2025). However, long-range rovers operating across the lunar day-night cycle may experience extended periods of low illumination in addition to traverses near permanently shadowed regions, unless they explicitly avoid those areas. These constraints motivate navigation approaches that do not depend on continuous high-quality imagery. In this work, we utilize a minimal navigation infrastructure that is agnostic to the illumination or imagery quality of the region.
Several studies have investigated rover or surface user positioning with limited lunar infrastructure. Tanaka et al. (2025) propose a minimalist system consisting of a single satellite, a known reference station, and a stationary user. They rely on differential dual one-way ranging for faster positioning convergence in comparison to Doppler measurements. Their study focuses on relative positioning for a stationary user and assumes dual one-way ranging capability, which is distinct from the one-way Doppler signal considered in this work for the space segment. Furthermore, we evaluate our method on a continuously moving rover. Jun et al. (2024) introduce joint Doppler and ranging (JDR) and demonstrate real-time estimation of position, velocity, and timing for lunar surface users using one-way range and Doppler measurements from two orbiters along with a surface reference station. Their simulation study assumed that the reference station’s clock was synchronized, which reduces emphasis on the network’s clock noise. In contrast, the architecture considered in this work assumes only a single satellite and does not assume that the relay broadcasts a dedicated ranging code. We also explicitly study non-ideal clocks at the rover, lander, and satellite nodes.
In prior work by the authors, we evaluated the localization performance of a rover using only the Doppler shift of a communication signal from a single relay satellite and demonstrated convergence from an initial 100 m 3D position error to sub-10-m accuracy within two relay orbits (Coimbra et al., 2025). In the present work, we incorporate the measurements that are available from the surface lander that deploys the rover, and we introduce a lander-aided differential Doppler correction. We also conduct an observability analysis to quantify the limited and unique geometry of the system.
PROPOSED WORK
This work evaluates the rover localization performance under a minimal navigation infrastructure concept. The mission scenario assumes a single relay satellite operating in an elliptical lunar frozen orbit modeled after the proposed orbit of ESA’s Lunar Pathfinder. We assume the lander position is known through human-in-the-loop map matching or pre-mission localization within 5 m 3D standard deviation.
The rover receives one-way Doppler measurements from the relay satellite downlink signal, as we assume that the satellite does not carry a dedicated navigation payload. The rover also receives one-way ranging measurements from the surface lander. The lander is equipped with a receiver that measures the Doppler shift of the relay downlink signal and broadcasts a differential Doppler correction to the rover. The correction is intended to reduce the sensitivity of the rover’s Doppler measurements to ephemeris errors and satellite clock effects without requiring the rover to estimate the full satellite state or requiring the lander to solve a standalone orbit determination problem. The rover uses only these two external measurement sources, together with onboard inertial odometry, to estimate its position over time using an Extended Kalman Filter (EKF). The resulting measurement link configuration is shown in Figure 1. We evaluate performance across several mission cases that isolate the effects of the lander ranging link and of the lander broadcast Doppler corrections. We also perform Monte Carlo evaluations to quantify the performance across stochastic realizations of the measurement and process noise.

Previous Abstract Return to Session G2a