DynamicVLA: A Vision-Language-Action Model for Dynamic Object Manipulation - Explained Simply | ArXiv Explained