Restaurant
Description
The robot retrieves and serves orders to several customers in a real restaurant previously unknown to the robot.
Main Goal: Detect calling or waving customer, reach a customer’s table without prior guidance/training. Take and serve all orders.
Optional Goals: Use an unattached tray to transport the order.
Task Breakdown
- The robot should identify a customer calling or waving.
- The robot should reach the customer’s table.
- The robot should take the order (keep eye contact).
- The robot should go back to the Kitchen bar and pick the requested object.
- The robot should return to the identified customer and serve the order.
Possible points
| Category | Action | Score |
|---|---|---|
| Regular Rewards | Detect calling or waving customer | 2×100 |
| Reach a customer’s table without prior guidance/training | 2×100 | |
| Take an order. | 2×300 | |
| Serve an order. | 2×300 | |
| Bonus Rewards | Use an unattached tray to transport | 2×200 |
| Regular Penalties | Not making eye-contact when taking an order | 2×–80 |
| Not reaching the bar (barman has to move from behind the bar to interact with the robot) | 2×–80 | |
| Deus ex Machina Penalties | Being guided to a table | 2×–200 |
| Asking the Barman to handover object to the robot | 4×–50 | |
| Guest needing to take the object from a tray or the robot’s hand | 4×–50 | |
| Being told/pointed where is a table/Kitchen-bar | 2×–100 | |
| Special Penalties & Bonuses | Not attending (see sec. 3.9.1) | –500 |
| Using alternative start signal (see sec. 3.7.8) | –100 | |
| Outstanding performance (see sec. 3.9.3) | 200 | |
| Total | Total Score (excluding special penalties & standard bonuses) | 2000 |
Tasks per Area
Vision
- Detect calling or waving customer
- Re-identify the person in case they lower their hand and the robot loses sight of them
- Detect objects from the kitchen bar
Navigation
- Reach a customer’s table in an unknown environment
- Return to the kitchen bar
- If computer vision is used to re-identify the customer, the robot should return to the customer’s table according to the customer’s position provided by the vision system. However, a different approach would be to keep track of the customer’s position by online mapping or other means.
Manipulation
- Follow the face of a person when taking an order
- Pick up an object from the kitchen bar
- Place the object on the table
HRI
- Take an order from a customer
- Communicate any needs to the bar man
- Interpret the objects to be served
Main states
Detecting customers
- In this state, the robot should look for a customer calling or waving.
Navigation to the customer
- The robot should navigate to the customer’s table.
- This could be achieved with a follow person approach as long as the customer can be re-identified.
- Another approach would involve online mapping.
Taking the order
- For this state, the robot should take the order from the customer.
Navigation to the kitchen bar
- The robot should return to the kitchen bar to pick up the objects.
Picking objects
- If a tray is used, the robot should request the objects from the barman to be placed in the tray or pick them and placing them in the tray by itself.
- If no tray is used, the robot should pick up the objects one by one.
Returning to the customer
- The robot should return to the customer’s table to serve the order.
Placing objects
- The robot should place the objects on the table.
- If a tray is used, the robot should serve the objects from the tray or request the customer to take them.
