Paper Reading #19: Local danger warnings for drivers: the effect of modality and level of assistance on driver reaction

Comments

Evin Schuchardt

Cindy Skach

Reference Information

Title: Local danger warnings for drivers: the effect of modality and level of assistance on driver reaction

Authors: Yujia Cao  University of Twente, Enschede, Netherlands
              Angela Mahr  German Research Center for Artificial Intelligence, Saarbrücken, Germany
              Sandro Castronovo  German Research Center for Artificial Intelligence, Saarbrücken, Germany
             Mariët Theune  Universtiy of Twente, Enschede, Netherlands
            Christoph Stahl  University of the Saarland, Saarbrücken, Germany
             Christian A. Müller  German Research Center for Artificial Intelligence, Saarbrücken, Germany

Presentation Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces ;
February 7-10, 2009;

Hong Kong, China

Summary

In this paper the authors discuss Advanced Driver Assistance Systems, and their use today. The user interface (the warning presentation), as the authors explain is "crucial to a successful danger avoidance". The authors have decided to perform a user study to investigate different types of warning presentations. This is mainly done by using a scenario of emergent road obstacles. The authors relate warning presentations with situation awarenes (SA) theory. Situation Awareness, the authors explain, has three hierarchical phases for a driver: perception of elements in the environment, interpretation and comprehension of the elements, and finally predicton of the future state of the environment.

The researchers considered two presentation factors: Modality, and Level of Assistance. Modality had 4 variants: speech warning, visual and speech warning with blinking cue, and visual warning with sound cue. The level of assistance varied with the presence or absence of Action Suggestions(i.e: change lanes’). For their experiment, the authors used simulation sofware that they put in a Mercedes-Benz. Scenes were projected of a highway with two lanes, with no "surrounding" vehicles. Visual warnings were displayed on a screen to the right of the steering wheel, and audio was fed through a speaker.  

The authors used different types of obstacles like fallen trees, rocks, lost cargo, or a broken-down vehicle appearing. These were described to the driver with the following parameters: type, location, distance from obstacle. Some setups they tested were:

Discussion

This paper presented interseting ideas. I was curious to know how one could test such a system, however the authors have managed a thorough experiment. The findings from the authors were interesting, and they corresponded to my predictions. As drivers often listen to music, I had a hard time seeing how to incorporate an auditory warning signal without disrupting a driver's listening experience. The researchers plan in the future to use a head-down display rather than a projected displayed on a vehicle's windshield.

• Visual warning with a blinking red bar
• Visual warning with an auditory signal
• Visual warning supplemented with speech
• Speech-only warnings

Results: Overall, the researchers found that adding warnings enhanced drivers' reaction time. For example, auditory warning aided in braking faster. As far as feedback from users, they expressed how they liked the assisstance provided, and found it useful. The authors also explained that a visual warning worked the best on drivers, and auditory warnings less so. The authors do explain, however, that a survery performed after their experiment revealed a preference for a beep rather than a bliking red bar, and that a visual warning without sound wasn't very popular.

Paper Reading #18: Speeding pointing in tiled widgets: understanding the effects of target expansion and misprediction

Comments
Shena Hoffman
Evin Schuchardt

Reference Information

Title: Speeding pointing in tiled widgets: understanding the effects of target expansion and misprediction

Authors:  Jaime Ruiz  University of Waterloo, Waterloo, ON, Canada 
               Edward Lank  University of Waterloo, Waterloo, ON, Canada 

        

Presentation Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces ;
ACM

New York, NY, USA ©2010

Summary

In this paper, the authors discuss a pointing facilitation technique, in addition to their desire to improve the time it takes a user to reach his/her pointing target when using a graphic interface. In particular, they wish to remedy tiled-like target arrangements via widget target expansion. We can give the example of the font and point size selection widgets, common in word processing programs like Microsoft Word.  A common problem that researchers face is the limited amount of screen space around tiled targets. Expansion of a target results in less available surface, and affects neighboring targets. Occlusion and displacement, the authors say, are two problems that can occur when a target is expanded. In this paper, the researchers have relied on previous work for single target expansion, to guide their design. The technique the authors have focused on is a technique for predicting where a user will select in real time. It is called kinematic endpoint prediction (KEP). The way it works is that it “uses the motion profile of a user to define a region of interest on the display.”

The researchers have performed two experiments to test the possibility of using KEP for multiple selected target expansion. In their first experiment, they tested to see expanding a certain set of targets'. In addition target displacement was assessed. Target displacement is the result of targets near the expanded targets shifting their positions to make room for the expanded target. They presented users with a random set of pointing tasks with some tasks using expansion with a simulated predictor said to be highly accurate and some using no expansion. There were seven possible tiles to select.

In the second experiment they used a similar setup, however in this case, the aim was to use the KEP algorithm to predict which targets should be expanded. The target arrangement used on the screen was different because the entire computer screen was filled with targets rather than just displaying seven targets.

Some of the findings of the two experiments were that:

• Target expansion does improve pointing and selecting performance, but better when there is a small set of targets.
• A minimum requirement for a target technique to be useful is an accuracy greater than 56.5 percent.
• There is a limit to how much a target can shift when expanded. As Fitt's Law explains, increase in distance reduces the likelihood of a user pinpointing his/her desired target.

DiscussionI liked this paper. The authors are trying to improve graphic interfaces by diminishing the time it takes a user to select their desired target. Current programs such as Microsoft Word that use the Ribbon style display, have organized targets in an intuitive way, however the sheer amount of widgets, and options eventually gets tiresome, especially when you are looking for a new widget. The authors' experiments were thoroughly explained, and the data they gathered was well analyzed. The technique the authors are trying to improve is very important as this is a very common task performed by users of graphic interfaces; therefore improving pointing performance by even a small margin can greatly improve a user’s productivity with graphical user interfaces.

Paper Reading # 17: A code reuse interface for non-programmer middle school students

Comments

Cindy Skach
Evin Schuchardt

Reference Information

Title: A code reuse interface for non-programmer middle school students

Authors: Paul A. Gross  Washington University in St. Louis, St. Louis, MO, USA 
              Micah S. Herstand  Washington University in St. Louis, St. Louis, MO, USA 
              Jordana W. Hodges  The University of North Carolina at Charlotte, Charlotte, NC, USA 
              Caitlin L. Kelleher  Washington University in St. Louis, St. Louis, MO, USA 
           

Presentation Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces; 
ACM New York, NY, USA ©2010

Summary

In this paper, the authors discuss an interface that allows non programmers (i.e: middle school students) to reuse code without requiring prior programming knowledge.  The interface is implemented for the Looking Glass IDE; the next version of storytelling Alice. Alice was a freeware object-oriented educational programming language with an integrated development environment (IDE) developed in the late nineties.

The way the environment is initially set up is with stories upon which functionality can be added. They give the example of a virtual girl, Melley, whose task is to develop super strength and knock a house over. After finding a similar story from a library(i.e: an agent who jumps at an "evil" doctor and knocks him down.), the user has to "steal" the story's Actionscript which starts the reuse process. The action occurring is then recorded. Provided with helper tools, the user can select the relevant code that will be used in his/her own story. To integrate the new code within the current story, the user must specify the taget actors of the script. In the example given, Melley would carry out the agent's task, and the house would carry out the "evil" doctor role. Hence, when the new script is played, you see Melley jumping, and knocking down a house just like we wanted. To resolve the possible issue of mapping actions to objects of different sizes the user has to reference subparts by name and choose more subparts if there is no match. 

A user study was performed with 47 students. The grade pool was from sixth to eighth grade. The students were given a tutorial, and then examples. They also were asked for feedback after usage. Students worked in groups and showed each other their animations. This encouraged them to reuse animations. 77% of the actions used in the animation contained more than five lines of code. Finally, the students took a quiz at the end of the study. They found that students who tried tinkering with the given code, rather than copying it, performed well on the quizzes.

Discussion

This paper was interesting. I always regretted not being exposed to programming concepts before arriving to college. With regards to the aforementioned program, Alice, I had read, and seen videos about it before. I believe it stems from Carnegie Mellon, and is related to the late Randy Pausch's research, if my memory serves me right. I actually tried the Alice program about two years back and liked it, so I'm on board with this present research.

Full Blog: Coming of Age in Samoa



Author: Margaret Mead

Book: Coming of Age in Samoa

Summary

In this book, the author starts by giving us the historical and sociological context of her work. She then states the general state of the american youth in the United States in the post first World War era. She evokes the overall sense of unease, and the difficulty of youth in their transition through adolescence. Attempting to understand the origins of the problems, she offeres us her anthropological background, and the anthropological method which is the study of another less complex civilization to extrapolate an explanation. The civilization she decides to study are the Samoan people. In  particular the samoan adolescent girl.

She proposes to explore two main questions:

• Are the disturbances which vex our adolescents due to the nature of adolescence itself or to the civilization?
• Under differenct conditions does adolescence present a different  picture?

From then, she gives a detailed description, about her experience in Samoa, and what conclusions she draws from that experience.
The summaries of the chapters of the book can be found in my previous blog entries.  At end, she answers the questions she posed at the beginning, and we can draw from her answers, that adolescence does differ based on location, culture, and morals.

Discussion

I felt this was good exposure to an ethnographic work. When reading Mead's book, one clearly sees that she put a lot of thought into her methodology, and study approach. This reflects in the amount of detail that can be found in the book when it comes to the Samoan people. I admit I'm usually for less description with books, but given that it's part of an ethnography, I can understand the need to do a thorough study. The only thing I did not appreciate, nor understand, was the need to go into details with Samoan sex practices. That's a little invasive in my opinion. Apart from that, I'm not sure why this book invited a lot of criticism. In the context of HCI, the only thing I think Mead teaches us, is that as program designers, we sometimes need to take a step back and consider another perspective in order to understand the way we do things.

Paper Reading #16: Using fNIRS brain sensing in realistic HCI settings: experiments and guidelines

Comments

Cindy Skach
Erin Schuchardt

Reference Information

Title: Using fNIRS brain sensing in realistic HCI settings: experiments and guidelines

Authors:Erin Treacy Solovey  Tufts University, Medford, MA, USA 
             Audrey Girouard  Tufts University, Medford, MA, USA 
             Krysta Chauncey  Tufts University, Medford, MA, USA 
             Leanne M. Hirshfield  Tufts University, Medford, MA, USA 
             Angelo Sassaroli  Tufts University, Medford, MA, USA 
             Feng Zheng  Tufts University, Medford, MA, USA
             Sergio Fantini  Tufts University, Medford, MA, USA 
             Robert J.K. Jacob  Tufts University, Medford, MA, USA 

Presentation Venue: UIST 2009: 22nd annual ACM  symposium on User interface software and technology; Date: 2009;
Location: New York, NY, USA

Summary
This paper deals with Functional Near-infrared Spectroscopy (fNIRS). In a nutshell, fNRIS measures the oxygen content of blood in the brain. It sends near-infrared light through the skull to a depth of 1.5 - 3 cm. This light is absorbed by both oxygenated hemoglobin and un-oxygenated hemoglobin and reflected back to sensors located near the light source. The data received from the sensors creates a cognitive state, which describes which sections of the brain are active.

 

As background, the authors give us some of the brain imaging techniques in use today:

• Functional Magnetic Resonance Imaging (fMRI) which is very effective in neural imaging, however does not do a good job when it comes to head movement. The slightest head movement results in corrupted images. In addition, the use of magnetic fields forbids the use of metal objects near the head.

 

• Electroencephalography (EEG). It also falls short as a viable solution because it requires participants to put gel in their hair. Since fNIRS does not require gel and is portable, the researchers believe it is the best brain imaging technique to be used in HCI.

Some historical background given by the paper reveals that brain imaging devices have been traditionally designed for disabled users, many of which paralyzed to receive special training. For the uses of HCI, the aim is to implement some type of cognitive control system. This explains the need to make the current available technology portable, and catered to things like head movement. A series of experiments have been performed to test the effects of these different sources of interference called artifacts. They focused on reading brain activity while the subject did nothing, used the mouse, used the keyboard, moved its head, made facial expressions.

They found that using the mouse and keyboard produced little or no interference with the fNIRS reading, where as head movement and facial expressions made the readings incomprehensible.

Based on these findings during the experiments the team outlined a set of guideline for using fNIRS in HCI. Future improvements may include using a chin rest so major head movements can be avoided and adding an accelerometer to the device so that head movements can be detected.

Discussion

This paper gave me a good introduction to brain imaging though I am a Computer Engineering major. This is a multidisciplinary project, that presents some good methods for reducing noise in fNIRS data. As mentioned in the paper, issues like a person's head movement will need to be resolved though. I do feel, however, that the authors of this paper have done effective work in laying down the foundation for future work on fNIRS. Future work, from this point, should be interesting. I am eager to see if we will ever be able to control our computers solely from the mind.