Camera Control at the Edge with Language Models for Scene Understanding