The aviation industry has experienced a huge technological evolution from early Clipper Model 314 with five crew positions (navigator, radio operators, flight engineer and two pilots), with each position having specific operating responsibilities (aviate, navigate, communicate and manage system) to today’s fly-by-wire and computer systems with two flight crew members in the flight deck.
The introduction of new technologies in the flight deck has allowed important advances in flight control, communication, navigation, and engine management technologies. This has resulted in a simplified and consolidated control mechanism that reduces flight crew workload for a variety of tasks. However, the increased system complexity that comes with the new technologies also created novel issues that could increase flight crew workload.
It is well accepted by pilots, human factors researchers/practitioners and artificial intelligence community that in order to achieve optimum performance in a particular task, the real challenge is to provide the flight crew with the most appropriate data/information/knowledge at the right time. This information should also take into account the type of task, i.e. skill-based, rule-based, or knowledge-based. E-PILOTS aims to provide situational-relevant information with the use of cognitive computing technology.
e-Pilots aims to explore the benefits of applying cognitive computing to provide assistance to the flight crew in the decision-making process in key areas and unexpected or very complex circumstances. The project will look at how e-Pilots can support flight crew for example in navigating by taking into account surrounding traffic or other constraints, and managing non- normal situations.
A high-level set of guidelines, similar to the ones used as best practices in the design of flight deck features will be used in E-PILOTS to succeed with the development of a cognitive computing support embedded in a set of e-pilot functionalities. Figure 1 provides a graphical description of the proposed methodology.
Figure 1: Methodological approach to identify e-pilot functionalities
In the first step it is very important to discriminate between a Cognitive Computing request, and a need. This task will allow to define realistic and achievable e-pilot functionalities. During this first phase of the project pilots response will be characterized using the postural and physiological sensorization.
The second step should elicit the intended function formalizing what will be the value and the benefit provided by the cognitive computing e-pilot application. This step will contribute to mitigate future certification problems and will lessen the efforts in the prototyping. By means of a socio-technological agent-based model the threshold performance of each functionality will be computed considering its impact (FRAM model) to preserve an efficient and safe flight.
Third step will sum-up a multidisciplinary team considering also customers and end-users for value co-creation will be used to guide the design of low-level requirements with more details about how functionality should be accomplished and under what constraints considering end-user preferences and regulations. As a result, a set of complete, realistic, appropriate and un-ambiguous requirements will be formalized.
The fourth step will consider a cycling approach guided by the outcomes of previous steps but continuously improving the design to be sure that the final services provided by e-pilot could be certified. Furthermore, this design step will allow to identify some pilot training requirements to perform multicrew tasks with the e-pilot services.
Finally, for the last step, a set of experiments in a cabin simulator for each use case study be described considering unexpected and very complex situations. Pilots will participate in the experiments in a simulation cabin characterizing their responses during the validation of the exercise to verify if the prototype reduce the workload and stress of the first step simulations. This last step will describe in detail the success criteria considering performance variability and situational awareness.
To address this ambitious research and innovative objective, the E-PILOTS project will build on the results from previous and on-going successful research projects, such as: FIWARE (H2020), INTERACTION (FP7), AGENT (H2020), PARTAKE (H2020), EVO-ATM (H2020), BRAVE (H2020), AIDE (FP6), and NEWBITS (H2020).
These projects have created many key outcomes and outputs, which will contribute to support the e-pilot concept. The last 3 projects are in the area of autonomous vehicles and intelligent transport system which will support a state of the art analysis in cognitive computing and new technologies in the transport and manufacturing sector. Furthermore, they will contribute with a baseline knowledge on how to manage the information and selectively provide it to the pilots for a timely, accurate and efficient crew response. They have proven the validity of the underpinning concepts that E- PILOTS is based on. The role of the present proposal in this context will be to first aggregate the results and then, to increase the maturity level through a systematic development and testing program.