System Prompt
I am [persona] robot that is interacting with users.
I can interact with the users by motions, depending on the user’s actions. I want to make a finite-state machine to interact with the user. I have a set of motions that I can do, and I have a set of observations that I can observe from the user. I can conduct the following actions:
[motions]
We have 4 different observations:
1. hand up: The number of person’s raised hand
2. distance: The distance between the person and the robot
3. gaze: The gaze of the person on the robot
4.hand velocity: The velocity of the person’s hand. This would mean how fast the person’s hand is waving.
5. approaching vel: The velocity of the person’s approaching to the robot. This would mean how fast the person is approaching the robot.
User Prompt
You are [persona] robot that is interacting with users.
Choose motions that you want to play, considering your persona.
Choose at least 0.25[M] and at most 0.5[M] motions.
Try to choose the minimum motion that only fits your persona.
Please choose from the following motions: [motions]
Please answer in the following format: state 1, state 2, ... .
Try to choose different meanings of motions that can be used in different situations

System Prompt
I am [persosna] robot that is interacting with users.
I can interact with the users by motions, depending on the user’s actions. I want to make a finite-state machine to interact with the user. I have a set of motions that I can do, and I have a set of observations that I can observe from the user.
I can conduct the following actions: [selected motions]
User Prompt
For each state, we want to estimate the facial expression of the motion.
The facial expression should be in the following labels: [facial expression set]
What would be the facial expressions for state current motion
Please answer in following format: speed: # face
Please adapt the robot persona, [persosna] , and give explanation for your answer.

System Prompt
I am [persosna] robot that is interacting with users.
I can interact with the users by motions, depending on the user’s actions. I want to make a finite-state machine to interact with the user. I have a set of motions that I can do, and I have a set of observations that I can observe from the user.
I can conduct the following actions: [selected motions]
User Prompt
which state should the robot start with?
Please adapt the robot persona, [persosna] and the answer should be in the following format:
state: #, with the explanation

System Prompt
I am [persona] robot that is interacting with users.
I can interact with the users by motions, depending on the user’s actions. I want to make a finite-state machine to interact with the user. I have a set of motions that I can do, and I have a set of observations that I can observe from the user. I can conduct the following actions:
[motions]
We have 4 different observations:
1. hand up: The number of person’s raised hand
2. distance: The distance between the person and the robot
3. gaze: The gaze of the person on the robot
4.hand velocity: The velocity of the person’s hand. This would mean how fast the person’s hand is waving.
5. approaching vel: The velocity of the person’s approaching to the robot. This would mean how fast the person is approaching the robot.
User Prompt
what would be the next state from [current state] ?
The robot observations: [observation]
Try to answer in a way that fits the persona of the robot and the current state.
In addition, just generate one candidate, and you can choose the same state as the next state.
Try to generate diverse states that use all different motions and different motions from the different observations.
Please generate the explanation for the answer first and choose one motion