SafeExit4All: An Inclusive Indoor Emergency Evacuation System for People With Disabilities
Seyed Ali Cheraghi, Anup Sharma, Vinod Namboodiri, Güler Arsal · 2019 · Proceedings of the 16th International Web for All Conference (W4A) · doi:10.1145/3315002.3317569
Summary
This paper presents SafeExit4All, a smartphone-based emergency evacuation system designed to help people with disabilities independently find safe exits from indoor spaces during emergencies such as fires, tornadoes, and active shooter situations. The authors identify a critical gap: while accessible indoor wayfinding systems are emerging, none had addressed the specific requirements of emergency evacuation, where users must quickly find alternative routes under stress, potentially avoiding dynamically changing danger zones. Existing evacuation aids like directional sound evacuation (DSE) systems and active shooter detection systems are static, unable to personalize routes based on individual abilities, and generally designed for only one type of disability. SafeExit4All is built as an application layer on top of GuideBeacon, a BLE beacon-based indoor wayfinding system. The indoor space is modeled as a connected graph where vertices represent points of interest (at beacon locations) and edges represent paths with weights encoding distance, surface type, passage width, and number of turns. When an emergency alert is received, the system determines the user's current location via beacon proximity, identifies danger zones defined by building administrators, and computes a personalized shortest path to the safest accessible exit using a modified Dijkstra algorithm. The system provides multimodal turn-by-turn navigation instructions through audio, visual, and haptic feedback, with the interface adapting to user preferences — BVI users receive audio narration via TalkBack, sighted users see a graphical map interface, and all users receive vibration and audio cues for directional guidance.
Key findings
The system was evaluated with 10 participants across five disability categories: blind cane users, users with light perception only, users with partial vision, a blind user with a guide dog, a motorized wheelchair user, and a walker user, plus sighted controls. In scenarios without danger zones, SafeExit4All reduced evacuation times by an average of 39% and evacuation distances by 84% compared to unassisted evacuation. With danger zones present, evacuation times were reduced by 12% on average (lower due to sighted participants with usable vision benefiting less) and distances by 39%. Critically, without the app, participants entered danger zones multiple times during evacuation attempts — even the sighted participant strayed into the danger zone despite being able to read room numbers. With SafeExit4All, no participant entered a danger zone. The personalized routing algorithm successfully accounted for different mobility needs: wheelchair users were routed to exits with ramps, and path weights were adjusted based on measured movement speeds for different disability types (e.g., wheelchair users take 3.5 seconds for a 90-degree turn versus 2.6 seconds for ambulatory users). Subjective ratings averaged 9.2 out of 10, with participants highlighting the value of clear, stress-reducing guidance in unfamiliar environments.
Relevance
This research addresses a life-safety application of accessible indoor wayfinding that has significant real-world implications. Emergency evacuations represent one of the highest-stakes scenarios where inaccessible design can have fatal consequences for people with disabilities. The paper's approach of personalizing evacuation routes based on individual mobility characteristics, sensory abilities, and preferences provides a practical model for inclusive emergency planning. The graph-based modeling of indoor spaces with accessibility-relevant edge weights (surface type, passage width, turn counts) offers a reusable framework for any indoor navigation system seeking to accommodate diverse abilities. For organizations and building managers, the administration tool that allows defining danger zones and monitoring evacuee locations demonstrates how technology can augment emergency response procedures. The study's inclusion of participants across multiple disability types — rather than focusing on a single category — makes it one of the more inclusive evaluations in this space and highlights how different disabilities create different evacuation challenges that a one-size-fits-all approach cannot address.
Tags: indoor navigation · emergency evacuation · wayfinding · visual impairment · mobility impairment · cognitive impairment · BLE beacons · personalized routing · inclusive design · assistive technology