This paper evaluates communication methods for safety area violations in human-robot interactions with Autonomous Mobile Robots (AMRs) in industrial settings. The study compares auditory, visual, combined auditory-visual communication concepts to each other and baseline without explicit communication. The concepts are implemented on a real robot and used in a within-design study with 33 participants. Results reveal that visual cues are significantly more effective in terms of perceptibility and intelligibility than auditory cues. The combination of auditory and visual signals shows comparable perceived safety and comfort to visual signals alone but with a lower level of frustration than auditoryonly. Overall, trust in the robot significantly increases post-experiment. The paper concludes by highlighting the potential of integrating visual or combined audio-visual cues for communicating safety area violations in AMRs.