Experiments with 40 volunteers show that collaborative robots improve human perception when they adjust behavior based on facial expressions and context. Researchers used Visual Language Models to bridge the gap between visual cues and emotional responses. This integration allows machines to better synchronize with human partners during shared physical tasks.