User Test: Perceived Behavioral Motion
In the story, an important aspect of the White Rabbit was conveying anxiety. The objective of this user test is the acquisition of a motion pattern to elicit this behavior. Furthermore, the test was used to validate the information given by movement expert Roos van Berkel, stating that by using high acceleration and high deceleration, and the importance of the left/right rotation of the head would make the rabbit appear anxious.

The test subjects will be presented a total of 12 motion patterns with modulations in curvature; curved and linear, acceleration and deceleration. Due to time constrains and the difficulty to program accelerations in a sine wave pattern, the curved pattern will be shown with only a linear acceleration and linear deceleration. The straight linear pattern will be shown with (1) Linear, (2) Low-Low, (3) Low-High, (4) High-Low and (5) High-High acceleration/deceleration respectively. The 6 motion patterns mentioned above will be repeated with head rotation to evaluate the influence on the perceived behavior. This leads to 6 motion patterns without the head rotation and 6 motions patterns including the head rotation, adding up to a total of 12 motion patterns for the user test.

For the user test, Flash and ActionScript3.0 have been used to present and simulate the various motion patterns with its modulations in acceleration and deceleration. The reason to choose a digital media over a physical one is based on the fact the reliability of depicting the exact same motion pattern for each individual test person. Flash allows the motion patterns to be programmed and the modulations easier to be implemented. Furthermore, the user test can be designed in such a way that the user will be able to conduct the entire user test by itself; buttons to repeat and to proceed to the next motion pattern, and fill in the questions digitally, without having to guide and supervise the user during the test.


A Rabbit’s Persuasion
Based on previous user tests with the current Stage 1 installation, several design opportunities were available regarding the interaction between the visitor, rabbit and the space itself. How can we persuade the visitor to follow the rabbit without any verbal instructions? It was difficult to provoke curiosity in such a way that the visitor follow the rabbit directly to the rabbit hole. Visitors did notice the rabbit, but most ignored or did not perceive the rabbit as persuasive enough in order for them to follow, as the rabbit moved towards the rabbit hole. A design proposal was made based on making a minor change in the physical space; placement of a hedge, and redesigning the movement path algorithm of the rabbit. Here 2 scenarios will be presented; either the user follows the rabbit or decides not to. If the user follows the rabbit immediately the rabbit just simply continues moving towards the rabbit hole.
Visitor Follows Rabbit
Otherwise the rabbit moves to and hides behind the hedge. Here the White Rabbit tries to draw the attention of the user again by making short motions e.g. making quick turns or “popping out” of the hedge, takes a quick look around and goes back hiding behind the hedge. The purpose is to trigger more curiosity to a state where the user creates empathy and wants to take peak at the rabbit. When the subject approaches the hedge the rabbit will be triggered move to the rabbit hole.
Rabbit triggers curiosity to persuade


Perceived Behavior Motion of Anxiety
In order to make the rabbit more convincing, research was done on the aspect of the visitor’s perceived behavioral motion of anxiety when observing the rabbit. The Laban Movement Analysis was used to determine a motion pattern based on the Efforts of Quick Time and Direct (quick and direct motion). An example of this movement can be depicted by a fast waving motion which stops abruptly from left to right. In terms of the rabbit’s movement, these characteristics can be converted to fast acceleration and deceleration, linear movement direction and sharp turns to elicit anxious behavior. And according to Roos van Berkel, a choreographer and motion expert, the head and limb movements are crucial factors that will support the depiction of an anxious movement. This is actually the opposite of the current setting, where the movement pattern has a large curvature and no stops. A user test was conducted to explore and validate these movements using Flash and a guideline for anxious motion was found for the ALICE Rabbit. The recommendation and guideline for an anxious motion pattern for the rabbit would be a straight and direct motion with a high acceleration, sharp turns when changing directions and occasionally rotate the head from left to right during the stops at a turn.
Anxious Behavioral Motion


Autonomous Motion System
The idea to develop an autonomous motion system for the rabbit was based on prior tests where the rabbit was controlled by a human operator through camera feed using a remote control. The rabbit was difficult to control, as the camera was placed in an angle and the real distance between the rabbit and obstacles were difficult to perceive.
Rabbit & Target Points
For the autonomous motion and as a proof of concept, a tracking system was designed with a top-down camera using an algorithm using Processing that is able to distinguish the rabbit using color tracking and calculate the discrepancy of the distance and angle between the rabbit and a target point. The discrepancy in distance and angle is calculated through Pythagorean Theorem equation and the angle is determined using the tangent trigonometry equation respectively.
Distance & Angle Calculation
The algorithm then sends out commands to the rabbit (AdMoVeo robot) through Bluetooth connection to execute the actual motion. This way, the operator can set target points in which the rabbit will navigate to autonomously, creating a more smooth and convincing motion.
Autonomous Motion System Flow Chart

The video above was a pre-test for the motion and rotation of the AdMoVeo robot based on human input in the Processing interface. The working prototype did have some inaccuracies in the precision and actuation of the motors. Unfortunately, it’s not possible yet with the current state of the robot to determine this precisely, because there is no feedback mechanism attached to the wheels. Inaccuracies could also be decreased by reprogramming and moving the processing code to the robot itself, to create a better timing for the exact rotational turn and travelled distance through motor actuations. Finally, the inaccuracies can also be lowered by using a faster computer, a more accurate tracking method and a faster wireless transmission with lower latency compared to the XBee.