Multimodal Integration Challenges in Emotionally Expressive Child Avatars for Training Applications