Better Human-Machine Communication Could Help Avoid Accidents

This communication system so frequently used among humans should also be used to exchange communication with machines. In fact, this concept of human-machine communication is being used more and more in an increasingly “hyperconnected” digital world through Internet of Things (IoT), interactivity or mobility offered by smartphones or wearables.

Human-machine communication is constant in different sectors, such as automotive, tourism, industry, etc. but it was the aeronautical sector where different European companies have come together to work in this field. To think of it, a message incorrectly coded or decoded among them in a specific context may have fatal consequences, and some air traffic accidents might have been prevented if this type of misunderstanding between the crew and the control had been corrected.

HoliDes, Holistic Human factors and System Design of Adaptive Co-operative Human-Machine Systems (AdCoS) is born under the assumption that pilots and automated cabin systems need to communicate and act together in a cooperative and highly adaptable manner. Establishing correct communication, co-operation and mutual adaptability will guarantee messages are understood and therefore tasks are completed fluently while ensuring safety at all times.

In global terms we can say work is being carried out with different sensors and algorithms measuring the contextual status of human operators so that once data from selected parameters have been processed and analysed, may become part of a system or tool helping with decision making and therefore develop a capacity for “cooperative adaptation” between human and machines and vice versa.

To achieve this, work is being carried out in different technological specialisation areas among which the following can be highlighted: definition of human behaviour patterns and design of a tool for interference; unification and storage of information related to experiments; human factor modelling ontology and preparation of recommendations for a correct human-machine communication.

The first block - design of the tool called Pilot Pattern Classifier - mainly consists of using sensors to detect and measure those parameters which are relevant to the behaviour of a pilot in service which may help to define and customise predictive models of human behaviour in real time (fatigue, distractions, etc.). A series of devices or sensors fitted to the pilot to monitor brain waves, eye movement, heart rate and sweating as well as other biometric measures required are the responsibility of Universita degli Studi Suor Orsola Benincasa.

This and other information of high documentary value such as experiments carried out by Honeywell in flight simulators, interviews with pilots, questionnaires, etc. must be stored in an accessible and structured way to be effectively operational. This data reservoir referring to different experiments carried out is being developed by Tecnalia and is called EDA, Experiment Data Archive. We can consider this a big data tool.

In addition to this and by way of this project development, the terms necessary to ensure all these systems using the information understand each other must be standardised. In other words, an ontology for human factor modelling needs to be generated for systems to be interoperable. Tecnalia is involved in this area as a leading company.

The latest area of this project is the preparation of communication recommendations. User guides or handbooks contain recommendations for messages to be correctly conveyed between sender and receiver, i.e. human-machine. To do so, the way information is presented is taken into account. This means any kind of variable which may affect the way a message is understood, such as visual factors, interfaces, (information architecture, colours used, final look of user interface, etc.), acoustic factors (type of sounds used, volume) and touch factors (vibrations) are also considered.

In sum, technology and tools to build an emotional intelligence system based on co-operative human-machine communication are applied in HoliDes. Sensor and algorithm systems are used to measure the human factors in different contexts and scopes, thus taking the first steps towards the design of adaptive systems (AdCos-Adaptive Co-operative Human-Machine Systems).