Answers are grounded in this paper and cite their sources.
Simulation study demonstrates improved landing navigation accuracy during Mars descent, indicating enhanced sensor fusion resilience against dust and poor lighting.
Mars landing navigation faces harsh environmental challenges such as visual degradation, uneven illumination, and dust interference. Traditional visual-inertial navigation methods exhibit limitations in robustness and generalization. Building upon the Efficient Deep VIO framework, this paper introduces a multimodal attention mechanism to enhance deep feature interaction between visual and inertial data, thereby improving navigation stability and modality adaptability in complex environments. To validate the effectiveness of the proposed algorithm, a Mars landing dataset was generated using Unity, including synchronized images and IMU data. Experimental results demonstrate that the improved method exhibits strong adaptability and accuracy under Mars landing conditions.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: