Throughout this project our group had to alter the typical methodologies we learned in class. Below you will find the various non-traditional techniques that we used throughout the project.

Non-Robotic Contextual Inquiries
With efforts to conduct user research in every important aspect parallel to the ultimate MER domain in which our interface is to be used in, we mapped out a matrix with the different aspects or features that were important to study, and compared each project that we intended to observe, to see which areas we could cover, and which were yet lacking.  The projects we had initially enlisted for observation were all in the robotic domain, however, currently, robotics work is predominantly completely autonomous or else the other extreme: teleoperated.  Therefore, there was not much dynamic re-tasking involved in many of the projects; mixed-initiative was hard to find. 

As a result, we turned to brainstorming other contexts in which we might find re-tasking.  For example, managers in fast-paced environments which needed to constantly reschedule/re-task their staff based on the current situation.  Places we looked to were fast-food chains, hospital OR, and police emergency dispatch.  We also thought that game environments would provide an excellent source of strategic planning, and re-tasking.     

Improvisation of Task
In order to gain early understanding of the task of commanding a non-collocated robot without having to have an already built system and robot to explore with, our team recreated a scenario of the game, using people to play out the different parts of the system.  One person was the robot, one played the operator, and a third was the note-taker.  With this improvisation of task, we were able to determine questions and demands the interface would have to support, at an early phase in our project. 

Extreme Scenarios
For transitioning into the creative process of the design phase, we thought up extreme usage scenarios for our future interface.  These extreme scenarios were intended to stretch our minds and give us new ideas and insights into design ideas.  For example, extreme scenarios involved commanding 1 billion robots, or really cheap, dispensable robots.  Other scenarios involved, really smart robots, or a precious, and delicate robot such as a human baby.   Questions we asked ourselves were: what were the implications of these scenarios for our interface?  How could we support these scenarios?  What might we be otherwise taking for granted?

Game Simulation to Understand Contingencies and Strategies
At a critical point in our process, we needed to narrow our focus.  We had many questions as to how people thought about contingencies, how far they went into explicitly developing alternative plans, how many alternates might they consider, and how detailed.  We hypothesized that there were 6 contingency types that could be made, but wanted to verify the validity of our hypothesis, and wanted to see how often each of them was made, and if users could recognize the types if we showed it to them. 

We developed a game simulation that had this exact purpose, in addition to familiarizing ourselves with the game concepts, without the requirements of needing a real robot all of the time.  The game simulation automatically returned feedback to the user that occasionally had errors so that the user would have further reason to re-task their robot.  We ran this game simulation with hard-core game players!  These players had years of experience in strategy, and offered us extremely valuable feedback.

Cognitive-Walkthrough / User Study
In various situation it was impossible to follow a strict regiment of only performing only a contextual inquiry, observation, or interview. For this reason we often times had to combine these various user study techniques to gather the most data possible in the richest environments. These techniques were performed at both the Jet Propulsion Lab (JPL) as well as the Mars Desert Research Station (MDRS). This customized method was very useful and not too difficult to adjust to and perform when needed. However, the data gleamed from performing the mixture of techinques was extreamly useful and priceless in helping us perfect our design.