Title: AI-Enabled Accessible Parking Monitoring and Enforcement: A Comprehensive Review of Computer Vision, ALPR, IoT, and Edge Intelligence
Authors: Navdeep Kaur, Deepti Sharda, Ritika Bansal, Deeksha Gupta, Mandeep K. Chawla
Volume: 10
Issue: 9
Pages: 4-15
Publication Date: 2026/09/28
Abstract:
Accessible parking is a critical element of inclusive mobility because designated spaces are intended to reduce physical and logistical barriers for people with disabilities. However, unauthorized occupation, misuse of permits, inadequate monitoring, and intermittent manual enforcement can reduce the practical availability of these spaces. Recent progress in artificial intelligence (AI), computer vision, automatic license plate recognition (ALPR), optical character recognition (OCR), Internet of Things (IoT), edge computing, and cloud platforms creates an opportunity to transform accessible-parking management from periodic inspection into continuous, evidence-assisted monitoring. This comprehensive review examines the technological components required for intelligent accessible-parking monitoring, including accessible-space identification, vehicle detection, parking occupancy analysis, license-plate recognition, accessibility credential detection, multi-object tracking, authorization verification, temporal reasoning, IoT communication, edge/cloud deployment, privacy, security, and human supervision. Unlike conventional smart-parking solutions that primarily determine whether a space is occupied, accessible-parking enforcement requires the integration of spatial, visual, temporal, identity, and authorization evidence. The reviewed literature indicates that individual components such as parking occupancy detection, object detection, ALPR, OCR, tracking, and IoT architectures have reached substantial technical maturity, whereas complete accessible-parking enforcement remains comparatively less developed. Existing datasets are also fragmented: general parking datasets provide strong support for occupancy research but generally lack accessibility-credential and authorization annotations. Directly relevant work demonstrates the feasibility of combining vehicle, plate, and disability-badge recognition with authorization services, while recent aerial-imagery research demonstrates the possibility of mapping accessible parking at scale. Based on these observations, this review presents an evidence-centric architecture that connects accessible-space mapping, vehicle detection, tracking, plate recognition, credential analysis, authorization databases, temporal reasoning, confidence estimation, secure event storage, and human review. The review further identifies challenges involving small and occluded credentials, environmental variation, cross-site generalization, privacy, legal variation, uncertainty, and the absence of standardized end-to-end benchmarks. The paper concludes with research directions focused on multimodal datasets, privacy-preserving edge intelligence, uncertainty-aware decision making, explainability, cross-condition validation, and human-centered enforcement.