Human robot interaction

Human-Robot Interaction There is a problem with clarity in the question – I've been assigned a robot, and the question accepts that, but the question asks "for which you design your robot." The robot has already been designed, so the question is nonsensical. There is also the claim that terms in capital letters are defined in Looijie – this is also false. Those terms are used in that paper, and form the basis of it, but they are not explicitly defined in that paper.

The instructor clearly is not very good at English and this is going to put the paper at risk because of it. 1. The main objective for which Sanbot designed the Elf appears to be for "Hi-level (sic) scenarios in the hospitality, retail and public service…" The sentence is not completed – Sanbot needs a copywriter in the worst way. There is no other information about what the robot actually does. My application will be to give this robot function, in this case in a supermarket.

The Elf will be programmed to assist shoppers with basic questions about the location of goods or service areas within the store. It will do this by accessing both basic store layout information (i.e "where are the bathrooms?"), but also specific SKU information (i.e. "where are the cans of Spam?"). If a good is out of stock, the Elf will be able to tell the customer that as well.

The Elf is capable of voice interaction, and 3D dynamic perception, which allow it to interact with human shoppers and respond to 1-on-1 questions of a dynamic nature. The Elf's touchscreen can help shoppers to search for specific items, and then display the location of the item or shopping area (i.e. deli) on a map of the store. 2. A use case specifies a specific situation where the methods will be used to meet an objective (Looije , et al, 2017).

A common use case would be a customer who wishes to find an item in the supermarket. We know that each supermarket takes a unique view of merchandising, so a good that is in one section in one store might be in a completely different section of another store. The Elf can assist customers to find everything that they want. The Elf will have access to a store map, both with different departments and with different SKUs. A customer would present to the Elf and make the query, eg.

("I can't find the 10kg bags of basmati rice."). The Elf would be programmed to check a) if there are any in stock and b) where they are located within the store. The Elf would then respond both verbally ("10kg bags of basmati rice are located in the middle of aisle 7") and visually using the robot's touchscreen to show the location visually.

The entire transaction can be done on the touchscreen as well – the Elf can be programmed to interact in multiple common languages, to serve customers who do not understand the default language. 3. The Sanbot Elf is capable of multiple interaction patterns. It can respond to speech, but it also has a touch screen that can be used for interaction. The use case above describes the ability to use the touchscreen to handle non-default languages, or to serve the hearing impaired.

It can also be used if verbal commands are not working (for example if the customer has a thick accent the robot cannot understand). Interaction patterns are means by which the Elf will understand, respond and be understood. It will need to be programmed with verbal fluency in at least one language, preferably all of the major languages in the region, and can be programmed with a broader set of languages for visual, on the touchscreen.

The Elf will need to be able to understand what language is being presented to it. Further, the different goods and departments within the store will need to be listed in each language, so that the Elf can find the good (eg 10kg of basmati rice will need to be understandable if presented to the robot in any of its languages). 4. Dautenhahn (2007) outlined a conceptual space of human-robot interactions. This takes a triangle structure where the three points are robot-centered view, human-centered view and robot cognition-centered view.

A place in the middle, D, is known as socially intelligent, and another space between robot and human-centered is known as socially situated. The Elf is basically a socially situated robot, which is defined as not needing to "possess any model of social intelligence", where interactions "emerge from the robot being situated in and responding to its environment." Interactions in this design are fairly simple, where only a limited amount of intelligence and adaptability are required for the robot to thrive.

There needs to be a level of human interaction, otherwise the Elf is just a touchscreen with a anthropomorphic casing, but this interaction need not be sophisticated, and is wholly within its context. 5. One of the physical limitations of the Elf is with its locomotion, in particular the arms and hands. The Elf is, for example, incapable of communicating with sophisticated gestures that would make it more humanlike, let alone performing something like sign language.

Ideally, the Elf would have more sophisticated limbs that allow it to perform more realistic interactions. The current design, which looks like the maid in the Jetsons, does not really convey a sophisticated level of interaction. From a software side, the ability of the Elf to do things like remember people so that it would greet you by name if you had interacted with it previously and given it that information, would also be valuable, especially since this robot is designed to work in a customer service setting.

Assignment 2 The authors studied the impacts of HRI, using a small sample size to draw their conclusions. This tested robot-assisted learning, so there was a teacher, but the robot was the variable that was changed between the control group and the study group. There were two study groups and one control group, so class size was not different between the two groups. The authors found that the robot helped reduce anxiety. Several factors were not examined in this study – other potential causes of anxiety being a significant one.

Were the three groups of equivalent composition? A larger study size would have made it easier to trust that there were no material differences between the groups. Leite performed a study in which children wore a wireless sensor to capture their electrodermal data, as a means of capturing their reaction to the robot.

The Leite study begins with a claim that "One of the main challenges in child-robot interaction is to capture children's spontaneous and genuine perceptions of the robot." There is no citation for this, but rather a tangential citation to a concept is psychology. The methodology of this study is rooted in this logic. Belpaeme et al also studied child-robot interaction. They also found benefits to using robots as part of a tutoring strategy. It was also.