Levels of abstraction in human supervisory control teams

Abstract

This paper reports a study into the Levels of Abstraction Hierarchy (LOAH) in two energy distribution teams. The original proposition for the LOAH was that it depicted five levels of system representation, working from functional purpose through to physical form to determine causes of a malfunction, or from physical form to functional purpose to determine the purpose of system function. The LOAH has been widely used throughout human supervisory control research to explain individual behaviour. The focus of this research is on the application the LOAH to human supervisory control teams in semi-automated 'intelligent' systems. A series of interviews were conducted in two energy distribution companies. The results of the study suggest that people in the teams are predominately operating at different levels of system representation, depending upon their role.Managerial personnel work at functional purpose and abstract function levels whereas operational personnel work at physical function and physical form levels. It is argued that both types of personnel are part of the wider distributed problem solving system, which includes both people and technology.KEYWORDS: Levels of Abstraction Hierarchy, Teams, Human Supervisory Control 1 IntroductionThe research literature has put forward the Levels of Abstraction Hierarchy (LOAH) as a description of five different levels of system representation (Rasmussen, 1983;. Studies have shown that these levels can be used to represent the decision space which is utilised by individuals in performing aspects of their task, shifting between the levels where appropriate (Vicente, 1999). The most persuasive arguments have been made by knowledge theorists (see Goodstein et. al., 1988) and empirical researchers (see Vicente, 1997;1999). Vicente, in particular, has demonstrated how experimental participants are able to perform process control tasks more effectively if they are presented with both functional and physical information about the system. This represents both end of the decision spectrum. Rasmussen has argued that this is because people need to work 'top-down' when seeking the purpose of functional requirements and 'bottom-up' when seeking causes of system problems.Many of the theoretical concepts in process control emanated from Rasmussen's work throughout the eighties (Rasmussen, 1983; classification was developed to assist system designers in better understanding human variability. Rasmussen intended to assist designers in building better interfaces, concluding that if system representation were more compatible with the operator's mental processes, there was greater likelihood of reducing human error and improving overall system performance (Goodstein, Andersen & Olsen, 1988). The approach has already been used to examine the roles of members of a nuclear power plant control team during different phases of operation (Gualtieri et al, 2000) and the respective roles of surgeon's and anaesthetist's in medicine (Hajdukiewicz et al, 2001). This research seeks to extend the analy...

A formative approach to the strategies analysis phase of cognitive work analysis

“…As is evident, most reports use Work Domain Analysis, with fewer reports using Control Task Analysis, and even fewer reports using Strategies Analysis, Social Organisation and Cooperation Analysis or Worker Competencies Analysis. Benda and Sanderson (1998), Bisantz and Mazaeva (2009), Bisantz and Vicente (1994), Burns (2000), Burns et al (2003,2007), Vicente (1998, 2001), Jansson et al and Sanderson (1999Sanderson ( , 2001, Nguyen (2002), Stanton et al (2006), Torenvliet et al (2008), Vicente (1990Vicente ( , 1991Vicente ( , 2002, Vicente et al (1995), Vicente and Rasmussen (1990), Watson and Sanderson (1998), Wong et al (1998), Xiao et al (2008), Xu (2007), Zhou et al As various researchers have noted, one reason that Strategies Analysis is seldom used may be the lack of a satisfactory method for performing it that is well-integrated with the other phases of CWA (Vicente, 1999, Naikar 2006, Kilgore et al 2009, Cornelissen et al 2012. In this article, we seek to address this need by proposing a method for formative Strategies Analysis that should apply to a range of domains.…”

Systematic literature review on work domains represented by the abstraction hierarchy

“…La catégorie du médical couvre des domaines de travail qui portent sur le corps humain [14] et un système de médication à domicile [23]. Le domaine de travail de la catégorie d'automatisation est un appareil photo [25], le domaine de travail de la catégorie des désastres naturels est une ville en cas d'inondation [2], le domaine de travail de la catégorie du jeu est le football européen [18], le domaine de travail de la catégorie du militaire est un navire de guerre [3] et le domaine de travail de la catégorie des ressources humaines est la supervision d'équipes [29]. Functional purpose [1, 2, 4, 6, 7, 10, 12, 15, 16, 18-20, 22, 23, 25, 29, 31, 36, 37, 43-45] Functional goals [17] General purposes [28] Highest level goals [3] Overall purpose [21] Purposes [14] Abstract function [4, 6, 7, 10, 12, 15-17, 22, 25, 28, 29, 43-45] Abstract function and priority measures [1] Abstract purpose [19] Balances [14] Generic control functions [21] Priorities and constraints [3] Values & priority measures; Priorities/values [2,18,20,23,36,37] Priority measures [31] Generalized functions [3,4,6,7,10,12,15,16,22,25,29,[43][44][45] General functions…”

A systems approach to flood vulnerability

“…Rasmussen's AH enables any civil engineering system to be represented at any level of granularity (Stanton et al, 2006). As a method for exploring such systems it becomes possible to insert (or remove) concrete physical objects from the bottom of the hierarchy, and to analyse the effect in terms of higher-level intangible activities, effects, outcomes and end states.…”