Robotic systems rely on perception to understand and safely interact with the physical world. However, no single sensing modality can fully capture the complexity of real-world environments. Our research investigates how complementary sensory information can be integrated to improve robotic understanding, decision-making, and autonomy in complex tasks.