The field of human-robot interaction and navigation is rapidly advancing, with a focus on developing more intuitive, adaptive, and effective systems. Recent research has explored the use of large language models (LLMs) to improve human-robot collaboration, enabling robots to better understand and respond to user instructions. Additionally, there is a growing interest in developing multimodal systems that can process and integrate multiple forms of feedback, such as visual, auditory, and haptic cues, to enhance navigation and interaction.

Notable developments include the creation of simulated environments for evaluating planning and navigation systems, as well as the design of frameworks that can generate executable code for multi-robot systems from natural language mission descriptions. Furthermore, researchers are investigating the application of LLMs in various domains, including aquaculture, warehouse management, and assistive robotics.

Particularly noteworthy papers include:

• NavVI, which presents a novel multimodal guidance simulator for visually impaired navigation in warehouse environments.
• OpenNav, which enables robots to interpret and decompose complex language instructions for open-world navigation tasks.
• Moving Out, which introduces a new human-AI collaboration benchmark that accounts for physical attributes and constraints in collaborative tasks.