Future planetary surface missions are likely to involve multiple independently operated rovers sharing the same deployment region, raising the need to verify compliance with operational constraints such as Lunar Safety Zones. Because continuous in-situ observability is rarely available, such verification requires post-hoc reconstruction of rover trajectories from sparse telemetry, including odometry, pose priors, and relative inter-rover detections. We introduce the problem of forensic trajectory analysis for non-cooperative planetary rovers in the presence of Byzantine agents: rovers that provide miscalibrated or deliberately falsified measurements to support an incorrect trajectory. We show that standard outlier-robust pose graph optimisation methods are vulnerable in this setting, because Byzantine rovers can generate measurements that are internally consistent and numerous enough to make truthful incriminating measurements appear as outliers. To address this, we propose an attribution-aware trajectory estimation method that reasons over rover credibility rather than individual measurement validity. The method evaluates candidate credible rover subsets by comparing the statistical consistency of their internal and boundary relative detections against provided priors, and then estimates trajectories using only measurements attributed to credible agents. Across synthetic simulations and real planetary-analogue trajectory data, the proposed method identifies Byzantine rovers and produces significantly more accurate trajectory estimates than existing robust pose graph optimisation baselines.
Figures & tables
Fig. 1: A high-level overview of our method, showcasing an example of credible rovers A and B with a Byzantine rover C. a) The trajectories of the rovers and relative detections between them, including the purported trajectory by C intending to hide its incursion into the Lunar Safety Zone. b) The likelihoods of relative measurements at a given time compared to pose priors, where misleading measurements from C have high likelihood, while incriminating measurements from A and B have low likelihood. c) Finding the optimal set of credible rovers by determining the set with jointly highest internal consistency and lowest boundary consistency.
Fig. 2: Distance and orientation errors at each timestep for credible and non-credible rovers by method in the case of a single Byzantine rover.
Fig. 3: A breakdown of distance errors at each timestep of non-credible rovers only by prior measurement kind {relative, absolute}, front and rear field of view (top), and by type of Byzantine behaviour and number of Byzantine agents (bottom).
Method
Model
Recall (%)
Precision (%)
1 Bz.
2 Bz.
Total
1 Bz.
2 Bz.
Total
Ours
RB
90.8
89.2
90.0
86.7
94.2
90.4
B
83.3
64.6
74.0
80.8
71.7
76.2
GNC
RB
7.1
4.4
5.7
20.1
13.5
16.8
B
4.2
5.0
4.6
19.8
18.3
19.0
PCM
RBO
6.4
5.4
5.9
31.1
45.2
38.1
TABLE I: Recall and precision of non-credible measurements by method, detection model, and number of Byzantine agents.
Fig. 4: Qualitative comparison of true trajectory, purported trajectory, and trajectory results for a single Byzantine rover in one of the simulation runs.
Fig. 5: Distance and orientation errors at each timestep for credible and non-credible rovers by method in the case of two Byzantine rovers.
Fig. 6: Distance errors at each timestep for non-credible rovers by method in the case of increasing relative measurement dropout rates.
Method
Model
Recall (%)
Precision (%)
Drop.:
0.0
0.5
0.8
0.95
0.0
0.5
0.8
0.95
Ours
RB
83.8
81.2
81.2
66.7
85.0
81.2
85.0
58.8
B
77.5
68.8
67.9
50.0
85.0
73.8
71.8
50.0
GNC
RB
8.5
4.3
2.1
0.2
27.5
18.6
18.5
2.5
B
5.7
7.6
1.5
0.0
33.0
32.4
14.5
0.0
PCM
RBO
7.6
4.6
6.3
16.7
30.3
26.4
20.6
20.0
TABLE II: Recall and precision by method, detection model, and dropout rate.
Fig. 7: The trajectories of three rovers A, B, and C, and the purported trajectory of Byzantine rover B avoiding the Safety Zone marked in orange (top); and a comparison of the trajectory estimates produced by different methods (bottom).
Lunar surface exploration increasingly relies on autonomous rovers that must perceive and traverse unstructured terrain under extreme illumination conditions. Near the lunar poles, low sun elevations produce long, high-contrast shadows that challenge stereo vision, terrain segmentation, and localization. Developing perception and navigation algorithms for these conditions requires large amounts of sensor data with accurate ground truth, which field campaigns and analog facilities can only provide at high cost and with limited variety. This paper presents MARE (Modular Autonomy and Rendering Environment), an open-source simulation suite for lunar rover perception and autonomy built on Unreal Engine 5 and ROS 2. MARE consists of three integrated components. Terrain Studio generates cratered and rocky lunar Digital Elevation Maps (DEMs) to build simulation scenes. A dedicated Unreal Engine project hosts two physics-based rovers with plugins for stereo, lidar, event, and Single-Photon Avalanche Diode (SPAD) cameras, streaming time-stamped sensor data and pixel-accurate ground truth. Lastly, a Guidance, Navigation, and Control (GNC) stack provides web-based mission control. Validation shows a mean Intersection over Union (IoU) of 0.79 between predicted and rendered shadows at low sun elevations, and stereo depth analysis verifies decreased pixel coverage in shadowed regions relative to sunlit terrain.
Raúl Castilla-Arquillo, Miguel Olivares-Mendez
Space Robotics (SpaceR) Research Group, SnT, University of Luxembourg, Luxembourg
Robot-aided exploration of planetary surfaces is essential for understanding geologic processes, yet many scientifically valuable regions, such as Martian dunes and lunar craters, remain hazardous due to loose, deformable regolith. We present a scout-rover cooperation framework that expands safe access to such terrain using a hybrid team of legged and wheeled robots. In our approach, a high-mobility legged robot serves as a mobile scout, using proprioceptive leg-terrain interactions to estimate regolith strength during locomotion and construct spatially resolved terrain maps. These maps are integrated with rover locomotion models to estimate traversal risk and inform path planning. We validate the framework through analogue missions at the NASA Ames Lunar Simulant Testbed and the White Sands Dune Field. Experiments demonstrate (1) online terrain strength mapping from legged locomotion and (2) rover-specific traversal-risk estimation enabling safe navigation to scientific targets. Results show that scout-generated terrain maps reliably capture spatial variability and predict mobility failure modes, allowing risk-aware path planning that avoids hazardous regions. By combining embodied terrain sensing with heterogeneous rover cooperation, this framework enhances operational robustness and expands the reachable science workspace in deformable planetary environments.
Shipeng Liu, J. Diego Caporale, Yifeng Zhang +17
University of Southern California, Los Angeles, CA, USA. · University of Pennsylvania, Philadelphia, PA, USA. · NASA Marshall Space Flight Center, Huntsville, AL, USA. +2
Accurate position estimation is crucial for the successful implementation of future lunar landings using autonomous vehicles, especially in dangerous environments with sparse terrain features. In this paper, we propose a terrain relative navigation (TRN) algorithm combining our deep-learning crater detector, which was designed specifically for the NASA Crater Detection Challenge problem, and an Extended Kalman Filter (EKF). Our detector analyzes crater features from the monocular images acquired from orbit, and their matches with craters from a global database are identified via a Hungarian assignment approach followed by the consensus-based outliers removal method. The estimated measurements are then used to refine an EKF, where spacecraft pose estimation in the Lunar-Centered Lunar-Fixed (LCLF) frame of reference, augmented with altitude aiding information, constrains radial drift. The simulation results indicate that even if the spacecraft is off from its actual location up to 5 km, TRN could recover from this situation, achieving navigation error reduction to a few hundred meters. It should be noted that in order to maintain crater feature correspondences, it is important to match the image resolution and the scales within the scene to the detector training set distribution.
Batu Candan, Simone Servadio
Aerospace Engineering Department · Iowa State University · Ames, Iowa, USA