Carl Westin

6-2 Decision-making in ATC conflict resolution 117 6-2 Decision-making in ATC conflict resolution Automation design approaches typically incorporate end users early in the design process. Similarly, several automated decision aids for conflict resolution have been developed around controller elicited decision-making models. 10 These typically view decision-making as a hierarchical search process based on controllers’ use of problem-solving heuristics, often separating conflict detection 33, 61 and conflict resolution. 24, 29, 31 Examples include the Conflict Resolution Assistance (CORA), 35 the Cube model 132 and the associated COCOS (Controllers’ strategies integrated into a conflict resolution system) algorithm, 31 and the Intelligent System for Aircraft Conflict Resolution (ISAC). 131 Acquiring a detailed understanding of controllers’ core cognitive work and decision-making processes represents a great challenge for the development of per- sonalized CD&R decision aids. Discrepancies have lead to models and design frameworks that poorly suit controllers’ working methods. 19, 155 Inferring cognitive processes is problematic since controllers typically struggle to express their deci- sions and actions, 29, 88 a commonly reported trait among experts in general. 236, 245 The naturalistic decision-making paradigm argues that experts typically do not con- sider several options when making decisions. Rather, decisions reflect the first cred- ible solution conceived intuitively through a process of pattern-recognition using tacit knowledge and previous experience. 245 Analogously, researchers have identi- fied CD&R as a naturalistic decision-making process,. 87, 131, 246 Controllers are believed to develop a strategy and solution “library” that is ac- cessed when encountering a conflict. 40, 89, 247 They appear to rely on the first con- ceived strategy, especially during periods of high workload. 189 Furthermore, con- trollers have been shown to rely on decision-making heuristics, typically described as “rules of thumb,” for both identifying and solving conflicts. For example, con- flicts tend to be approached in pairs and sequentially rather than globally if multiple aircraft are involved. 29 However, controllers have been found to rarely interfere with both conflicting aircraft. 88 Furthermore, although objectively both aircraft involved are in conflict with each other, one aircraft is typically assigned the “trouble maker” causing the conflict. 29 In addition, controllers have been found to avoid speed for solving conflicts en-route. 29, 31 Kirwan and Flynn 29, 35 suggest that resolution strategies primarily are shaped by four aspects. Factors represents contextual parameters such as phase of flight, conflict parameters, and remaining distance to destination. Rules encompass rules of the air and other airspace constraints. “No-no’s” represents valid solutions that controllers refrain from such as vectoring an aircraft in front of another, vectoring it 180 degrees, or climbing it when close to its destination. Principles represent general guidance strategies such as maintaining fairness, avoiding knock-on effects,