Mental Model Consensus and Shifts During Navigation System-Assisted Route Planning

Abstract

A major barrier to effective spatial decision-making in human-agent teams is that humans and algorithms use different mechanisms to solve spatial problems, frequently leading them to produce different solutions. Incongruity between algorithm-generated solutions and human spatial mental models results in higher workload in mixed-initiative systems, and potential breakdowns in trust and team situation awareness. Although these performance effects are well-understood, few methods exist for quantifying and comparing human spatial mental models and algorithm-generated solutions. To address these problems, 27 participants completed solutions to 5 spatial planning problems, before and after receiving assistance from 2 navigation algorithms. A novel path mapping and clustering approach provided a means of quantifying consensus in human mental models, and shifts in those mental models after viewing the algorithm-suggested routes. Human solutions clustered into a small number of shared mental models. Individual differences in trust in each algorithm predicted acceptance of that algorithm’s route.

Human–Agent Team Dynamics: A Review and Future Research Opportunities

“…Hou's IMPACT model [95] delineates six pivotal elements-Intention, Measurability, Predictability, Agility, Communication, and Transparency-that underpin human trust in IAs. Empirical studies corroborate that trust enhances the congruence of mental models between humans and IAs [96].…”

Where Do You Think You're Going?

“…These aforementioned studies of human variability did not explicitly examine convergence, however. Based on prior work using similar methodology (References [7,10], which found such effects when humans planned routes for themselves or a UAV under supervisory control, respectively), we expect that participants will exhibit this convergence when planning a route for themselves, as well as when planning a route for the driverless vehicle.…”

“…The result of this analysis was a distribution of values for each cluster, for each criterion, on each map. 7 These resultant distributions were compared using MANOVA to determine if each cluster's routes differed meaningfully on these criteria. For each of the 20 clustered MDS solutions, the MANOVA found significant multivariate effects indicating significant differences in the spatial mental models with respect to the aforementioned criteria (all p < 0.001).…”

Integrating Context into Artificial Intelligence: Research from the Robotics Collaborative Technology Alliance

“…The main difficulty is that humans are very quick to infer meaning from their surroundings. However, the process that robots use for inferring meaning from the world often does not match the process or capabilities of the human, leading to different planning and reason‐based decision making (Perelman et al 2017; Perelman et al 2018; Schaefer et al 2018). Although difference is not necessarily negative, it does affect the development of shared situation awareness and shared understanding among the team (Chen et al 2014; Wright et al 2017).…”