The Accreditation Board of Engineering and Technology (ABET) has declared that to achieve accredited status, 'engineering programs must demonstrate that their graduates have an understanding of professional and ethical responsibility.' Many engineering professors struggle to integrate this required ethics instruction in technical classes and projects because of the lack of a formalized ethics-in-design approach. However, one methodology developed in human-computer interaction research, the Value-Sensitive Design approach, can serve as an engineering education tool which bridges the gap between design and ethics for many engineering disciplines. The three major components of Value-Sensitive Design, conceptual, technical, and empirical, exemplified through a case study which focuses on the development of a command and control supervisory interface for a military cruise missile.

In time-critical systems in which a human operator supervises multiple semiautomated tasks, failure of the operator to focus attention on high-priority tasks in a timely manner can lower the effectiveness of the system and potentially result in catastrophic consequences. These systems must integrate computerbased technologies that help the human operator place attention on the right tasks at the right times to be successful. One way to assist the operator in this process is to compute where the operator's attention should be focused and then use this computation to influence the operator's behavior. In this paper, we analyze the ability of a particular modeling method to make such computations for effective attention allocation in human-multiple-robot systems. Our results demonstrate that it is not sufficient to simply compute and dictate how operators should allocate their attention. Rather, in stochastic domains, where small changes in either the endogenous or exogenous environment can dramatically affect model fidelity, model predictions should guide rather than dictate operator attentional resources so that operators can effectively exercise their judgment and experience.

Performance degradation due to interruptions is a critical issue, particularly when people are supervising highly autonomous systems in time and safety critical environments. Previous research in the development of automated support to help supervisory control operators resume task activities after an interruption has had limited success. This paper describes two new interruption recovery approaches that attempt to mitigate the disadvantages of previous approaches. In particular, this paper describes the design and initial investigation of a prototype assistive interface developed to support interruption recovery for supervisory control of multiple unmanned aerial vehicles (UAVs). These two 'replay'-type interruption recovery approaches enable increased user-control of the event discovery process and provide event 'bookmarks' to highlight emergent system events. The findings from this initial study provide several recommendations for the future design of interruption assistance tools for human supervisory control tasks.

In the near future, large, complex, time-critical missions, such as disaster relief, will likely require multiple unmanned vehicle (UV) operators, each controlling multiple vehicles, to combine their efforts as a team. However, is the effort of the team equal to the sum of the operator's individual efforts? To help answer this question, a discrete event simulation model of a team of human operators, each performing supervisory control of multiple unmanned vehicles, was developed. The model consists of exogenous and internal inputs, operator servers, and a task allocation mechanism that disseminates events to the operators according to the team structure and state of the system. To generate the data necessary for model building and validation, an experimental test-bed was developed where teams of three operators controlled multiple UVs by using a simulated ground control station software interface. The team structure and interarrival time of exogenous events were both varied in a 2x2 full factorial design to gather data on the impact on system performance that occurs as a result of changing both exogenous and internal inputs. From the data that was gathered, the model was able to replicate the empirical results within a 95% confidence interval for all four treatments, however more empirical data is needed to build confidence in the model's predictive ability.