Analysis of 1933 papers submitted to IROS 2026 reveals that while large language models (LLMs) are increasingly integrated into robotics research, they are not replacing traditional robotics disciplines. Instead, LLMs are being interwoven with established areas like robot learning, perception, planning, control, and manipulation, leading to a more complex, systems-oriented approach. Key trends include a focus on improving the efficiency and real-world applicability of LLM-powered robots, addressing challenges in 3D spatial understanding, long-term memory, and adaptability to changing environments. The research is shifting from simply scaling up models to building robust systems that integrate LLMs with other robotics components. AI
IMPACT LLMs are being integrated into robotics, shifting focus from pure model scaling to system engineering for efficiency and real-world application.
RANK_REASON Analysis of academic papers from a major robotics conference. [lever_c_demoted from research: ic=1 ai=1.0]
- Control / Dynamics
- Embodied AI
- GeoVLA
- Humanoid / Legged
- IROS 2026
- Jorge Mendez-Mendez
- LLM
- manipulation
- Navigation / Planning
- Origin Reality Lingji
- PDDLStream
- Perception / Vision
- robot learning
- Tianjin University
- Tsinghua University
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
