I am working on a self-developed project and I need some help on finding the right prompt or prompts to create a video.

I am trying to display the London Underground during rush hour. The commuters are on a descending escalator. The screen to their left is a ‘heat-sensored’ screen - so when it senses a body near it, a branch is displayed. The branch moves with the commuter down the escalator - and has a very gentle motion - as if swaying form a gentle breeze. It is important to show that this is heat sensored screen - so the spacing between the people and branches much be exact. The video should also show commuters entering the scene - and a branch appearing - again to show the relationship between commuter and branch. 

I used the native Kling application with free credits and I got a good visual result - but the relationship between the commuter and branch was not right. I have since payed for subscription on Firefly and used multiple programs (including Kling 3.0) I have tried multiple prompts and am not getting a good response at all. I have also tried breaking down the instructions - ie creating a video of people going down the escalator - before introducing a secondary prompt that introduces the branches, but with no luck. 

Attached is also my initial kling video generation - The people are obviously wrongly facing the screen but this was due to the reference image.

Any help would be greatly appreciated.

Thank you!