Please add support for long-horizon simulation tasks ：Mobile navigation、Obstacle avoidance for both dynamic and static obstacles during navigation , execute VLA (Vision-Language-Action) model-based operations

[moshilangzi]

Please add support for long-horizon simulation tasks that integrate the following capabilities:

1. Mobile navigation to reach target destinations.
2. Obstacle avoidance for both dynamic and static obstacles during navigation.
3. Upon reaching the destination, the agent should be able to execute VLA (Vision-Language-Action) model-based operations.

This would enable comprehensive evaluation of embodied AI agents in realistic, sequential task settings.