DRAWBOT: Making Everyday Objects Interactive

In various public settings, such as coffee shops, individuals often experience boredom while waiting in lines, a problem exacerbated by the lack of engaging and interactive experiences. This paper introduces DRAWBOT: Drawing Robotic Assistant for the Withering of Boredom or Tediousness. The project aims to address the challenge of uneventful waiting times by deploying a roboticized furniture system that encourages artistic expression in people.


INTRODUCTION
In an era dominated by technological advancement, perhaps "Making Everyday Objects Interactive" will assist in converting routine events into captivating and unforgettable moments.What are compelling responses to the challenge of enhancing everyday encounters?While previous work in robotic furniture demonstrates people's ability to anthropomorphize the motions of minimal robot constructions (see 2.1), and other eforts have sought to use machines to augment human creativity (see 2.2), this efort combines the two: ofering roboticized furniture to support human creativity.
Our frst stab into this heretofore underexplored domain was to utilize the Pioneer 3DX robot, augmented with a barstool for height and drawing surface.DRAWBOT utilizes its unique bar stool ziptied in place design and expert human wizards to approach bored humans and encourage them to participate in artistic expression.DRAWBOT's unique design and human-in-loop control enabled it to approach customers, with wizards seeking to encourage people in the frst foor of an engineering center near a cafe, study area, and classrooms to draw.
Its deployment took place over a four hour period close to Halloween, and improvizationally explored playful approaches, rotations while people were drawing, or slowly moving away to keep people on their toes.Our ethics board approved analysis of robotic and human behavior based on this data collection, which compiled insights from the three robot drivers, video-analysis, and observations about the drawings themselves (7 pages and 30 unique drawing instances).Concepts like being in the participants view, and choosing people who had time to spare seemed to increase drawing engagement over the course of the exploratory study.
Overall, we fnd that, yes, the people waiting-for-food were *most* likely to engage with DRAWBOT, but also that many people just-passing-though the atrium often paused to engage DRAWBOT as well.Interestingly, not a single person standing-in-line participated in drawing (perhaps not wanting to lose their spot).identify prior eforts to roboticize furniture and use technology to support human creativity as relevant to this work.

Robot Furniture
Within the feld of social robotics research, robot furniture refers to furniture that functions as a minimum social robot, ofering a lowcost platform that may be easily tested in public areas.A wriggling trash can [4] indicate that it wants someone else's trash, for instance, while an ottoman approaching someone's feet could be taken as an ofer [10].Social reactions can be elicited by environmental robots, like an automated door [8], drawers that open and close [7], and persuasive robot chairs [1].In these situations, people can anticipate the behaviour that the object may demand from them thanks to the functional category of the item.For example, a nonanthropomorphic robot ottoman in [10] used the robot's approach and jiggling to gave the impression that it was demanding their participation.This resulted in participants feeling prompted to put up their feet.

Technology aided creativity
In recent times, there has been a boom in AI-generated art.An artist could use text-to-image platforms as a tool [2] for creating and generating creative products such as poetry, stories, jokes, music, paintings, and so on [9] [6].
AI can also function as both a creative tool and an active collaborator when co-creating material with a user thanks to advances in Machine learning.Co-creation between humans and AI has recently been studied in music composition [5] or design ideation [3], to help artists get over their creative blocks.

TECHNOLOGY IMPLEMENTATION
This section provides a description of the hardware in section 4.1, and of the human-in-loop control software in section 4.2.

DRAWBOT Assembly
DRAWBOT is constructed of a drawing area atop a barstool atop a mobile robot (Fig. 2), so that people can draw without stooping down.A Pioneer 3DX is a 2 wheeled mobile robotic platform with a caster wheel using diferential drive.The stool is attached to this base using zip-ties on the cross bars of the stool to the handles on the Pioneer.A white cloth drapes the stool, tied to the robot handles.This allowed the height of the drawing surface to be a comfortable height of 65 cms.
The drawing supplies and instructions were placed on the drawing surface (Fig. 3).The drawing surface had a surface area of 30cmx30cm with a taped 8.5"x11" sheet of white paper, and a small bin (as one might use to make banana bread) flled with drawing tools, like crayons.A sign was attached to the side of the robot revealing study details and DRAWBOT description, "Use the Crayons to Draw something Fun."

Human-in-Loop Control Software
DRAWBOT utilized a velocity controller in the Robotic Operating System (ROS) framework.A wireless gaming controller was used to get input from a human wizard.One joystick was mapped to the linear velocity and another to the angular velocity.DRAWBOT was

STUDY OVERVIEW
The study unfolded over a 4-hour period, commencing shortly after noon on a Friday, coinciding with the approaching Halloween festivities, on October 27, 2023.The chosen venue was the atrium of a Computer Science building, encompassing study zones, tables, chairs, and a cofeehouse that extends across the frst foor.This setting, frequented by a diverse mix of individuals including students, faculty, staf, tour groups, and visitors, ofered a dynamic environment.

Setup
Throughout the entire deployment, DRAWBOT was operated by a human controller seated among the chairs near the cofee shop.The controller's primary objective was to cultivate a sense of creative interaction among cofeehouse patrons.
The observed participant cohort during the study comprised three primary groups: individuals awaiting their turn in the cofee shop queue, those positioned nearby anticipating their orders, and passersby.[Fig.6] After the initial hour of operation, a spontaneous incentive was introduced to enhance participant engagement: candy was added to DRAWBOT's tray, a decision fttingly made in the spirit of the ongoing holiday festivities.
The controller observed the participants' reactions and we placed two cameras to capture the scene.The study's comprehensive data collection approach included the systematic collection of drawings created by participants, meticulous observations and self-reports from the human controller, and the video record capturing the entirety of the scene.

RESULTS AND OBSERVATIONS
Over a 4-hour period, 7 sheets of A4 paper were drawn upon (Fig. 5, we subselected to 6 for space purposes) and 61 people interacted with the system.We analyzed the drawings, wizards observations, and the video data of 25 unique participants, wherein the criteria were a good video angle and actual drawing on the paper,

Drawings
We counted 58 unique drawing instances and a thematic analysis of the content created by the participants unveils a general positive response to DRAWBOT.Participants not only manifested their interests through doodles of nature, cartoon characters, and vehicles, but also expressed support for sports teams through slogans, demonstrated academic inclinations via mathematical formulas, and contributed positive afrmations through encouraging words.

Wizard Observations
Wizards reported trends related to human subgroupings in the space, efective robot behaviors, and approach strategies.• People waiting in line to order: Wizards reported that people standing in line did not draw on the robot, though they were approached several times.Wizards would try and move the DRAWBOT as the line would progress or remain stationary, while only rotating ever so slightly.Instead, drawing activities predominantly occurred from those waiting, and also (at a lower rate) from those passing by.• People waiting to collect their order: Wizards observed that once the participant has placed their order and is waiting to collect it, they would be more inclined to interact with DRAWBOT.In this scenario, wizards would simply go up to the participants and stop right in front of them, enabling the participants to interact with the DRAWBOT in many ways.• Passersby and others: For the people just passing by in the atrium, various strategies were employed to get them to interact with DRAWBOT.DRAWBOT would walk alongside them, approach them from the front and stop a few distance away from them, or be in their line of motion and wiggling/twitching slightly as the participant comes nearby.The wizards reported that out of these strategies and approaches the most efective was approaching a participant from the front, stopping a few distance away from their line of motion and twitching slightly.These behaviour patterns made the DRAWBOT seem alive and the people were keen to interact with it.
A common theme observed across all the groups was that people would stop interacting with DRAWBOT if they knew that someone was controlling it.The coded participant behaviors around the robot presented in Fig 7 are consistent with these wizard observations.

Video Analysis
We analyzed all the videos from cameras recorded during the duration of the study.In total, about 116 people passed through the frame of our overhead video: 39 were customers of the cafe, and 77 were passing by over the four hour period (more interactions occurred out of frame, that the wizards were able to report on).On average people spend approximately 14 seconds interacting with DRAWBOT.We found the following type of people's reaction to the DRAWBOT Note that (3)-( 7) are not mutually exclusive and (3) is inclusive of ( 4)- (7).In Fig. 7 we only used (3) We found these typical 7 cases as described above.Then, we summarized (4) - (7) into "(3)Interacted with DRAWBOT" because these cases were not mutually exclusive and presented them in Fig 7

CONCLUSION
This project involved a four hour public deployment of a mobile drawing platform that we dubbed DRAWBOT, as it could assist people to draw, while they battled boredom.We report on the observations of the three robot drivers who assisted in this study, on the 61 unique drawings collected across seven pieces of behavior, and provide video stats about the durations and means of drawings.
Since we observed a preference for DRAWBOT to be autonomous, we would work to make that possible in the future.Our goal would be to develop autonomous systems that utilize some of the existing wizarded approach strategies.This will enable us to codify further useful interactive behaviors that DRAWBOT and other robots can employ to communicate with humans who would fnd value in their services.Also, in further attempts, we would conduct a comprehensive survey to better gauge the participant's responses and recommendations.
Overall, we found that many people were willing to engage with the robot.

Figure 2 :
Figure 2: Break Down of DRAWBOT with the intersection points marked in red

Figure 3 :
Figure 3: The top of DRAWBOT showing the drawing surface, drawing supplies, and the controllers used by wizards.

Figure 5 :
Figure 5: Drawings Gathered during the Study Fig 6 displays common participant groups in the space.On one side of the cafe, they congregate into a line, those waiting for the order distribute in the open space, and passersby typically pass close to the tables on the opposite side from the cafe.

Figure 6 :
Figure 6: An example of subgroups observed during the Study, where A denotes people waiting for their order, B denotes passersby and C denotes people waiting in line to order.

( 1 )
Unnoticed: People passed by without noticing the robot (Fig 4(a)).83 People showed this behaviour, out of which 30 were part of a college tour.(2) Ignoring: People passed near the robot and saw the robot but did not stop walking(Fig 4(b)).7 People exhibited this behaviour.(3) Interacting with DRAWBOT: People passed near the robot and stopped to interact.

Figure 7 :
Figure 7: Participants Behaviour Coded from Video