Journal article
Cognitive and behavioral markers for human detection error in AI-assisted bridge inspection
Applied ergonomics, Vol.121, 104346
11/2024
DOI: 10.1016/j.apergo.2024.104346
Abstract
Integrating Artificial Intelligence (AI) and drone technology into bridge inspections offers numerous advantages, including increased efficiency and enhanced safety. However, it is essential to recognize that this integration changes the cognitive ergonomics of the inspection task. Gaining a deeper understanding of how humans process information and behave when collaborating with drones and AI systems is necessary for designing and implementing effective AI-assisted inspection drones. To further understand human-drone-AI intricate dynamics, an experiment was conducted in which participants’ biometric and behavioral data were collected during a simulated drone-enabled bridge inspection under two conditions: with an 80% accurate AI assistance and with no AI assistance. Results indicate that cognitive and behavioral factors, including vigilance, cognitive processing intensity, gaze patterns, and visual scanning efficiency can influence inspectors' performance respectively in either condition. This highlights the importance of designing inspection protocols, drones and AI systems based on a comprehensive understanding of the cognitive processes required in each condition to prevent cognitive overload and minimize errors. We also remark on the visual scanning and gaze patterns associated with a higher chance of missing critical information in each condition, insights that inspectors can use to enhance their inspection performance.
Details
- Title: Subtitle
- Cognitive and behavioral markers for human detection error in AI-assisted bridge inspection
- Creators
- Fatemeh Dalilian - University of IowaDavid Nembhard - Harvey Mudd College
- Resource Type
- Journal article
- Publication Details
- Applied ergonomics, Vol.121, 104346
- Publisher
- Elsevier Ltd
- DOI
- 10.1016/j.apergo.2024.104346
- ISSN
- 0003-6870
- eISSN
- 1872-9126
- Language
- English
- Date published
- 11/2024
- Academic Unit
- Business Analytics
- Record Identifier
- 9984658252302771
Metrics
1 Record Views