Interaction Design and User Experience in Autonomous Vehicle Interfaces A Systematic Literature Review
Description
Dataset Description This dataset supports a PRISMA guided systematic review examining whether interior interaction design and UX research for fully autonomous (SAE Level 4–5) vehicles is fragmented and lab dominated. Three linked files trace the pipeline: RAW_DATA is the original Scopus export (479 records); PROCESS_DATA records three-stage PRISMA screening, reconciling 479 records to 85 included studies; FINAL_PROCESS_DATA contains the per-study extraction table (Appendix A) plus all descriptive statistics and figures reported in the manuscript. Findings- trust and usability dominate studied UX factors, evaluation remains simulator-dominated, real-world and longitudinal evidence is scarce. Use files in this numbered order to reproduce all PRISMA counts and statistics. How the three files connect: 1. RAW_DATA (SLR SCOPUS-AUTONOMOUS_INTERACTI, KEY WORDS, PRISMA sheets) the unmodified Scopus export, 479 records, before any screening. 2. PROCESS_DATA (Screening_Data, PRISMA_Summary, Key_Stats sheets) — applies the three-stage screening logic (title/abstract → interior-design relevance → automation-level) to reduce 479 → 85, with a live PRISMA summary. 3. FINAL_PROCESS_DATA (Included_Studies_Data, STATISTICS, APPENDIX A sheets) the 85 included studies with all coded fields (RQ mapping, methodology tags), formula-driven descriptive statistics (means, medians, IQRs matching Tables 2–5 and Figures 2–4 in the manuscript), and the manuscript's Appendix A extraction table.
Files
Steps to reproduce
Methods, Protocols, and Workflow Data were gathered via a PRISMA 2020-guided systematic search of Scopus (single database) on 02 July 2025, using the Boolean string: ("autonomous vehicle" OR "self-driving car" OR "automated vehicle") AND ("interaction design" OR "UX" OR "user experience" OR "HMI" OR "human–machine interface"), restricted to English-language journal articles, 2016–2025. Screening followed three sequential stages, each manually coded by the review author directly in the spreadsheet: (1) title/abstract relevance against inclusion/exclusion criteria (479→214) (2) interior interaction-design relevance (214→165) (3) vehicle automation-level check, retaining only SAE Level 4–5 studies (165→85). No automation tools or second-reviewer inter-rater check were used (disclosed as a limitation in the manuscript). For the 85 included studies, descriptive classification fields (AV/service context, methodology, UX/HMI focus, RQ mapping) were generated by keyword/phrase matching against each study's title, author keywords, and abstract using Python (pandas), not full-text human coding. Descriptive statistics (mean, median, SD, interquartile range via the QUARTILE.INC method) were computed with native Excel formulas in the FINAL_PROCESS_DATA workbook, so every reported figure is live and auditable, not hard-coded.
Institutions
- Indian Institute of Technology HyderabadTelangana, Sangāreddi