Natural language processing in robotic systems has become central to the development of intelligent autonomous agents capable of seamless human–robot interaction. However, the predominant focus on high-resource languages, particularly English, limits the global applicability and inclusivity of robotic communication systems. This research proposes a novel framework that leverages cross-lingual data science methodologies to enable robust robotic communication capabilities across English and Uzbek linguistic domains, incorporating specialized components for multilingual understanding, action grounding, cultural adaptation, and bidirectional learning. Experimental validation demonstrates that the proposed approach achieves task success rates of 89.3% for English commands and 87.8% for Uzbek commands, representing a significant advancement over existing methodologies. The performance gap between languages was minimized to just 1.5%, compared to disparities exceeding 5.6% in baseline systems. The integrated cognitive architecture demonstrated superior capability in translating linguistic commands into executable actions, with the action grounding module achieving a 91.2% success rate in command-to-action conversion. Cultural adaptation mechanisms contributed to an 18.7% improvement in communication naturalness ratings from Uzbek-speaking participants, while the morphology-aware representations enabled a 23.4% improvement in command-understanding accuracy for complex Uzbek structures. Ablation studies reveal that cultural adaptation contributes most significantly to user satisfaction metrics, while action grounding has a substantial impact on technical performance measures. The findings demonstrate that carefully designed cross-lingual frameworks can effectively bridge linguistic divides while maintaining the precision required for robotic task execution in real-world environments. This work contributes to the theoretical understanding and practical implementation of inclusive cognitive embodied intelligence systems, enabling more accessible and effective human–robot interaction across linguistic boundaries. The proposed methods can be extended to other low-resource language scenarios, supporting more equitable global deployment of robotic technologies.