Thursday, May 31, 2012

Using Games for Training

On the flight from Michigan to Minneapolis last week, I came across an article in Delta Sky magazine which described how corporations such as Chick-fil-A, Cargill, and Cold Stone Creamery are using simulations to train their employees. For example, Cold Stone Creamery uses a game-based simulation in which employees get points based on their ability to serve the right portion size and the speed with which they serve customers. Simulators are also used extensively in safety-critical domains such as medicine, military, aviation, space, and nuclear power to train key skills to operators.

Ever wondered what the fidelity of a simulator should be? In very simple terms, fidelity refers to the extent to which a simulation mimics the real world. Fidelity includes equipment, audio-visual, physical, motion, and cognitive fidelity.

Flight simulator

The big question is to what extent should a simulator mimic the real world so as to transfer skills to the real world?

The following line of research would imply that higher the simulation fidelity, the better the transfer.
  • Transfer of training occurs between tasks only if the tasks are similar or have common task elements. In Thorndike and Woodworth’s classic study, very little transfer was found when participants were asked to estimate areas of triangles, rectangles, or irregular figures after being trained on estimating areas of rectangles of a different size.
  • Skill retention is believed to be enhanced by increasing the similarity between the learning and retention environments.
  • Learning is context-dependent. In the classic Godden and Baddeley study, divers who learned a list of words on land and underwater recalled better when both learning and recall took place in the same environment.

Contradicting the above line of research, there is also evidence that suggest that a simulator does not have to replicate the real world in it’s entirely. That is, as long as the training and transfer environment share the same cognitive processing strategies and goals, skill transfer will occur. In fact, if all training was indeed context-specific, will we be able to apply anything that we learn in everyday life?

The following steps will help guide the design of a simulator.
  • Start with a cognitive task analysis of the procedure that you want to train operators on. That is, break down the overall procedure into constituent steps.
  • Next, identify the criticality of each step in the procedure.
  • Incorporate performance metrics for each of the critical steps, in the simulator.
  • Finally, establish the construct validity of the simulator by showing that the simulator is able to distinguish the performance levels of experts and novices.

Photo credit Bruce McVicar via Wikimedia Commons.

Monday, May 21, 2012

Designing catheters for cardiac procedures

Cardiac catheterization lab

Catheters are used in a myriad of cardiac catheterization procedures. In these procedures, a catheter is inserted into the femoral vein or artery through an incision in the groin and then advanced into the heart. The catheter is then used to complete the necessary procedure, such as deploy a stent or a valve. The catheter is removed from the body after completing the procedure.

Catheterization procedures eliminate the need for open heart surgery and offer various advantages such as faster implant time, reduced hospitalization, and faster recovery.  Some human factors considerations when designing catheters for such procedures are summarized below.

  • During catheterization procedures, physicians are fixated on the X-ray or fluoroscopic imagery. Fluoroscopy is an imaging technique that allows physicians to obtain real-time images of the heart structures during a procedure.
  • Physicians manipulate the catheter (using their hands) and look for feedback on the X-ray image. Because virtually all the visual resources of the physicians are dedicated to the imagery, they rarely look down at the catheter during a procedure.  So if the catheter has multiple buttons or controls on it representing multiple functions, these must be easily distinguishable to physicians by touch (i.e., tactile differentiation).
  • Designers frequently think that being able to visually discriminate between controls (i.e., using color or labeling) would be sufficient. This is inadequate in catheter design and has the potential to create user errors and confusions, especially when there are multiple controls on the catheter each depicting a unique function.
  • It is advisable to use redundant coding mechanisms such as size, shape, texture, and location when designing catheter controls so that users are able to differentiate between the controls through touch.
  • Often during these procedures, one hand of the physician is at the access site (where the catheter is inserted) and the other hand is on the catheter handle – torquing the catheter, manipulating the controls on the catheter and so on. Moving the hand away from the access site is undesirable as this has the potential to compromise the stability of the catheter. Therefore, it is important that the controls in the catheter handle be designed so as to facilitate a one-handed operation.

Photo credit Vuk at the German language Wikipedia, via Wikimedia Commons. 

Thursday, May 17, 2012

Robots for the Elderly

The FABBS Foundation recently published an article on the design of robots to assist the elderly in their activities of daily living. This article shows how psychologists are working collaboratively with robotics engineers to understand the end users’ needs and expectations, and in this case the needs of a very special population (i.e., the elderly), and incorporating these into robot design.

Key takeaways from this article:
  • Older adults will use technology if they understand the benefits of the technology. This is contrary to the popular belief that older adults do not embrace technology.
  • Older adults do not want robots to assist them in tasks that they are capable of doing themselves.
  • Older adults want robots to help them in tasks involving privacy.
There are several factors that need to be taken into consideration when designing personal service robots. Some of these factors are summarized below:
  • What tasks should the robot accomplish? The robot’s capabilities must be designed by taking into account the user’s expectations. For example, the needs and expectations of the elderly population would be very different from those of parents seeking child care assistance from robots.
  • How will the robot know when the user needs assistance? On the flip side, how should users control the robot or communicate their intentions to the robot?
  • When is the appropriate time for the robot to interrupt a user?
  • How should the robot approach the user? Gender and social setting are important considerations.
  • What is a reasonable distance at which the robot should position itself from the user?
  • How should the robot capture the user’s attention and then communicate to the user?
  • How should the physical appearance of the robot be? There is evidence that users expect human-like robots to have more communication capabilities than machine-like robots.

Photo credit Jiuguang Wang via Wikimedia Commons.

Tuesday, May 15, 2012

Icon design

NPR’s Science Friday website has a new look. I was drawn to the use of icons on their webpage.  Topics such as planet, space, brain, biology, nature, and mathematics have been very creatively represented using icons.

Using icons in design have numerous advantages:
  • Icons can contribute to the simplicity of the design by reducing the amount of text and clutter.
  •  Icons can be comprehensible to a wider population irrespective of language barriers, including children.
  • Icons can promote faster recognition. Remember a picture is worth a 1000 words!

Though icons offer several advantages, these have the potential to create a lot of user confusion if not designed correctly. Poor icon design on displays can also lead to user errors in safety-critical domains.

What are some of the techniques that can be used to ensure that icons represent what the designer intends to convey to the end user?

·      Phrase generation procedure:  In this technique, icons are presented to test users one after the other. Icon comprehension is measured by asking users to generate as many phrases as possible that come to mind when they see an icon. The first phrase generated from the series of phrases would indicate the concept that was most readily activated when a participant sees an icon. This is an important factor to consider when designing for safety-critical domains, where users are expected to react quickly to events. Icon comprehension scores would help designers to infer the intuitiveness of the icon design. Details on this technique can be found here.

·     Usability testing: In this technique, icons are presented to users as part of a design prototype that is being evaluated. Test scenarios should be designed that would require users to interact with these icons. Objective measures such as scenario completion time, number of clicks, and number of deviations from the optimal navigational path can be used to determine the intuitiveness of the icons. This can also be supplemented with a think-aloud protocol that would require users to verbalize the intentions behind their actions. While the think-aloud technique offers the advantage of getting to know users’ thoughts and expectations, this has the potential to interfere with the objective data (i.e., scenario completion time). Retrospective think-aloud may be more effective, wherein users are asked to explain their actions after completing a scenario. 

Tuesday, May 8, 2012

Autonomous Cars

As someone who did her dissertation on human-automation interaction, this article on CNN certainly caught my eye. I wanted to dissect the issue and examine the pros and cons of driverless cars.


  • Reduce driver workload: In this day and age, any extra time is good. With a car that drives on its own, drivers can now offload their attention to other tasks.
  • More safety: Accidents caused as a result of driver distraction can be avoided as long as the system is reliable.
  • Help the impaired: Individuals with difficulty driving (e.g., the blind) can now easily travel around.
  • Reduce carbon footprint: The automated system would be better at maximizing fuel efficiency in comparison to a human driver.


  • Out of the loop syndrome: The catch as is stated in the article is that “though a driver is not needed, one is in the front seat to take control when needed”. So what happens when the system fails - Will the operator in the front seat be able to take over control? 
    • A lot of research in the human-automation interaction area, including my own work, has shown that fully automated systems are not desirable as these have the potential to leave the operator out of the decision-making loop. 
    • Operators become over-reliant on automated systems and fail to monitor the system. Several factors are responsible for this – high automation reliability, automation consistency, when operator’s confidence in the automation exceeds confidence in self and so on. 
    • The costs associated with overreliance on automation are unfortunately higher for higher degrees of automation.
The Federal Aviation Administration (FAA) also has plans to modernize the future airspace to account for the increasing traffic volume and this proposal involves a lot of automated systems in the cockpit and in the air traffic control facilities. With all these efforts going on in the road and in the sky, will the world be a safer place for humans? I am certainly not implying that we should stop making technological advancements but when doing so, the human performance consequences associated with these should not be forgotten and should be examined. 

Photo credit Flickr user Steve Jurvetson under the creative commons license.