Project topics and materials
Get Complete Project Material File(s) Now! »
Information Diffusion via Voice Communication
Many of the tasks presented in Section 2.2.1 are done through communicating with an aircraft or after communicating with an aircraft. In performance these tasks, a controller accepts inputs, process information, prioritizes, and acts(Rodgers and Drechsler 1993). Controllers’ response to these tasks, are all communication-related which are reflected in the controllers’ workload (Stein 1985). Before the slowly emerged data link communication, verbal communication was the only way for controller and pilot to exchange information. Recent work have demonstrate that controllers have clear preference for the data communication while pilots are reluctant to use data communication (Lacher, Battise et al. 2011). Therefore, verbal communication is likely to remain the prime means for controller-pilot communication for many years. Although there are many factors affecting controllers’ activities and consequently influence the system, from a system perspective, it is the controller’s voice communications that influence the system operation. By definition, the activity is a coherent system of internal processes and external behavior and motivation that are combined directed to achieve conscious goals (Bedny and Meister 1997). Thus we assume controller’s voice communication activity encapsulates both cognitive efforts and physical efforts to accomplish the mission of ensuring traffic safety and efficiency. Controller’s communication activity is referred as the event that controller press the push-to-talk button and hold in order to send the transmissions to aircraft. Normally the contents of the transmission should contain the information that to which aircraft the communication is addressed. It is very rare that one transmission includes more than two flight call-signs separated by ”break”. Hence, air traffic controllers’ communication activity with pilots can be seen as the information diffusion process.
External Activities: Voice Communication Activities and Performance
Before the slowly emerged data link communication, for instance the Controller Pilot Data Link Communication (CPDLC), voice communication was the only way for controller and pilot to exchange information. It is still the primary controller-pilot communication way in most of air traffic control centers. Analysis of air traffic controller voice communication data has a long history. In the past, communication events were extensively used to measure workload (Cardosi 1993; Manning, Mills et al. 2002; Manning, Fox et al. 2003; Coffey, Harrison et al. 2011; Lamb, Bartlett et al. 2011). Communication times (Corker, Gore et al. 2000), Communication durations are all found to be good measures of workload (Porterfield). The number of communications between controller and pilots were found significant related to both traffic volume and traffic complexity (Bruce, Freeberg et al.). Porterfield investigated the correlations between the controller’s communication duration of the 4 minutes prior to a rating and the controller’s subjective workload ratings. It was found that the correlation coefficient was 0.88 with p + 0.01(Porterfield). Rantanen et. al. investigated the impact of audio delay and pilot delay in air traffic controllers’ communication on controllers’ performance and workload(Rantanen, McCarley et al. 2004). Manning et al. have examined the relationship between communication events, subjective workload and objective task-load measures. The communication events used in their study were total number of communications, total time spent communicating, average time spent for an individual communication, and communication content. Although some measures of communication events are highly correlated with workload, the analysis indicates that voice communication metric does not make a unique contribution to the workload evaluation (Manning, Mills et al. 2002; Manning and Pfleiderer 2006).
Human Dynamics: Empirical Evidences
We used to assume that most of human actions occur randomly. The basic assumption of human dynamics models, which are used from communications to risk assessment, had been that the temporal characteristics of human activities could be approximated by Poisson processes. The difficulty of collecting experimental and real data had limited the quantitative investigation of human activity, which resulted in that the hypotheses and conclusions were given in qualitative. Thanks to the rapid development in electronic information technology, human activities data can be easily record which provides a perfect platform for studying human behavior. There is increasing evidences showing that the inter-event times, defined by the time difference between two consecutive activities, indeed follow non-Poisson statistical distribution. Heavy-tailed distributions of
inter-event times have been widely reporting from various kind of human activities, ranging from correspondence (Oliveira and Barabasi 2005), email communication (Malmgren, Stouffer et al. 2008; Malmgren, Hofman et al. 2009), through printing behavior (Harder and Paczuski 2006), online films rating (Zhou, Kiet et al. 2008), short message texting(Wu, Zhou et al. 2010), to human mobility (Gonzalez, Hidalgo et al. 2008). Instead of randomly occurring as assumed previously, the temporal patterns of human actions exhibit the bursts of frequent actions separated by long periods of inactivity. The similarities of the distribution of inter-activities times among human beings indicate that the way human do things is irrelevant to the contextual conditions. For example, Figure 3- 1 plots the distributions of response times for the letters replied by three famous scientists. In can be seen from the figure, all the inter-events times are well described by the Power Law form with exponent %1.5. In Appendix I give the summary of different human activities that have been analyzed.
Table of contents :
Abstract
Acknowledgements
Résumé
0.1 Contexte
0.1.1 Background
0.1.2 Énoncé du problème et la portée de la recherche
0.1.3 Motivation
0.1.4 Objectif de la recherche
0.1.5 Contributions
0.2 Les données de testes
0.2.1 D1 Dataset
0.2.2 Dataset
0.2.3 Dataset
0.2.4 Dataset
0.2.5 Dataset
0.3 Les caractéristiques temporelles des activités de communication des contrôleurs
0.3.1 Définitions
0.3.2 Les corrélations entre les activités des contrôleurs et la complexité du trafic aérien
0.3.3 DFA des activités des contrôleurs
0.3.4 Les temps inter-communication
0.3.4.1 Temps inter-arrivée
0.3.4.2 Longueurs des écarts inter-communication
0.4 Les caractéristiques spatiales des activités de communication des contrôleurs
0.4.1 Transformer les séries temporelles en réseau
0.4.1.1 Définition des noeuds
0.4.1.2 Détermination des arêtes
0.4.1.3 Réseau temporel agrégé
0.4.1.4 Réseaux temporels
0.4.2 Résultats
0.4.2.1 Distribution des degrés (Degree Distribution)
0.4.2.2 Les corrélations entre les réseaux communautaires et le trafic aérien
0.4.3 Réseaux temporels
0.4.3.1 Distribution des degrés dépendant du temps
0.4.3.2 Motifs du réseau
0.5 Fluctuation d’échelle des activités de communication des contrôleurs
0.5.1 Résultats empiriques
0.5.2 Modèle
0.6 Conclusions
Chapter 1 Introduction
1.1 Background
1.2 Problem Statement and Scope of Research
1.3 Motivation
1.4 Objective of Research
1.5 Contributions
1.6 Organization of the Thesis
Chapter 2 Air Traffic Control and Air Traffic Controllers’ Activities
2.1 Air Traffic Control (ATC) and ATC System
2.1.1 The Static (Physical) Part of ATC System
2.1.1.1 Airspace
2.1.1.2 Airport
2.1.1.3 Regulations
2.1.1.4 Software
2.1.2 The Dynamical Part of ATC System
2.1.2.1 Aircraft
2.1.2.2 Weather
2.1.3 Human Part of ATC System
2.1.4 Characteristics of ATC System
2.2 The Role of Air Traffic Controller
2.2.1 ATC Tasks
2.2.1.1 Separation tasks
2.2.1.2 Monitoring tasks
2.2.1.3 Constraint tasks
2.2.1.4 Coordinate tasks
2.2.1.5 Information tasks
2.2.1.6 Request tasks
2.2.1.7 Other tasks
2.2.2 Controller as a Black Box
2.2.2.1 Inputs of the controller
2.2.2.2 Outputs of the controller
2.2.3 Voice communication
2.2.3.1 Standard Phraseology
2.2.3.2 Contents of the ATC communications
2.2.4 Information Diffusion via Voice Communication
2.2.5 Summary
2.3 The State-of-the-Art on the Research of Air Traffic Controllers’ Activities
2.3.1 Tasks Demands: Air Traffic Complexity
2.3.1.1 Dynamics Density
2.3.1.2 Other Complexity Metrics
2.3.2 Internal Activities: Cognitive Activities and Workload
2.3.3 External Activities: Voice Communication Activities and Performance
2.3.4 Discussion
2.4 Chapter Summary
Chapter 3 The Temporal Characteristics of Controllers’ Communication Activities
3.1 Introduction
3.1.1 Human Dynamics: Empirical Evidences
3.1.2 Human Dynamics: Models
3.1.2.1 B-A Model
3.1.2.2 Cascade Poisson Model
3.1.2.3 Interaction Model
3.1.3 Comparison with Air Traffic Controller’s Activities
3.1.4 Objective
3.1.5 Definitions
3.2 Data
3.2.1 D1 Dataset
3.2.2 Dataset
3.2.3 Dataset
3.2.4 Dataset
3.2.5 Dataset
3.3 Correlations between Airspace Activities and Controllers’ Communication Activities
3.3.1 Communication Measurements
3.3.2 Dynamic density (DD)
3.3.3 Complexity based on dynamical system modeling (C-DSM)}
3.3.4 Results
3.3.4.1 Correlations between N and t C N t T
3.3.4.2 Correlations between communication, DD, and C-DSM
3.4 Temporal Characteristics of Controller’s Communication
3.4.1 Periodic Patterns of Controllers’ Communication
3.4.2 Detrended Fluctuation Analysis (DFA)
3.4.3 Inter-communication Times Distribution
3.4.3.1 Inter-arrival Times
3.4.3.2 Inter-communication Gap Lengths
3.5 Psychological Interpretation of the Intervals
3.6 Chapter Summary
Chapter 4 The Spatial Behavior of Controllers’ Communication Activities
4.1 Introduction
4.1.1 The Spatial Behavior of the Controllers
4.1.2 Motivations
4.1.3 Objectives
4.1.4 Related work
4.1.4.1 Structure-based Abstraction
4.1.4.2 Human Mobility
4.2 Method
4.2.1 A Network Approach
4.2.2 Mapping Time Series to a Network
Definition of the Nodes
Determination of the Edges
4.2.2.4 Temporal Networks
4.2.3 Network Analysis Techniques
4.2.3.1 Classic Techniques
4.2.3.2 Community Detection
4.2.3.3 Motifs Detection
4.3 Data
4.4 Results
4.4.1 Time Aggregated Networks
4.4.1.1 Degree Distribution
4.4.1.2 Correlations between Network Community and Air Traffic
4.4.2 Temporal networks
4.4.2.1 Time Dependent Degree Distribution
4.4.2.2 Network Motifs
4.5 Chapter Summary
Chapter 5 Fluctuation Scaling in the Controllers’ Communication
5.1 Introduction
5.2 Fluctuation Scaling
5.2.1 Temporal Fluctuation Scaling
5.2.2 Ensemble Fluctuation Scaling
5.3 Data
5.4 Results
5.4.1 Temporal Fluctuation Scaling
5.4.2 Ensemble Fluctuation Scaling
5.5 Model
5.6 Chapter Summary
Chapter 6 Implications of Air Traffic Controllers’ Dynamics
6.1 Implications for a Model-based Simulation for the ATM System
6.2 Implications for the Study of Cognitive Activities
6.2.1 Implications for the Resource Allocation
6.2.2 Implications for the Systems Design
6.3 Implications for other Human-Driven Complex Systems
Chapter 7 Conclusions and Perspectives
7.1 Summary
7.2 Perspectives
