Beyond Physical Reach: Comparing Head- and Cane-Mounted Cameras for Last-Mile Navigation by Blind Users

TL;DR

The paper addresses last-mile navigation challenges for blind users by systematically comparing head- and cane-mounted cameras. Through a formative survey of experienced cane users and a controlled study with synchronized head- and cane-mounted sensors across five real-world environments, it evaluates SLAM localization and NeRF-based scene reconstruction. Findings show head-mounted cameras deliver superior localization and mapping stability, while cane-mounted views offer broader ground-level coverage and richer reconstructions; a hybrid head+cane configuration yields the best overall performance. The work emphasizes designing with cane integration in mind and advocates adaptive, multisensor navigation systems co-designed with blind travelers to provide robust, user-aligned guidance.

Abstract

Blind individuals face persistent challenges in last-mile navigation, including locating entrances, identifying obstacles, and navigating complex or cluttered spaces. Although wearable cameras are increasingly used in assistive systems, there has been no systematic, vantage-focused comparison to guide their design. This paper addresses that gap through a two-part investigation. First, we surveyed ten experienced blind cane users, uncovering navigation strategies, pain points, and technology preferences. Participants stressed the importance of multi-sensory integration, destination-focused travel, and assistive tools that complement (rather than replace) the cane's tactile utility. Second, we conducted controlled data collection with a blind participant navigating five real-world environments using synchronized head- and cane-mounted cameras, isolating vantage placement as the primary variable. To assess how each vantage supports spatial perception, we evaluated SLAM performance (for localization and mapping) and NeRF-based 3D reconstruction (for downstream scene understanding). Head-mounted sensors delivered superior localization accuracy, while cane-mounted views offered broader ground-level coverage and richer environmental reconstructions. A combined (head+cane) configuration consistently outperformed both. These results highlight the complementary strengths of different sensor placements and offer actionable guidance for developing hybrid navigation aids that are perceptive, robust, and user-aligned.

Figures (3)

• Figure 1: Examples of key visual and tactile landmarks encountered during last-mile navigation tasks, including floor transitions, textured surfaces, doorways, tactile paving, and environmental boundaries. While this figure emphasizes features detectable via vision or touch, participants also described using auditory and olfactory cues.
• Figure 2: Semi-dense 3D point clouds and global closed-loop trajectories captured across five locations (Outdoor1–2, Indoor1–3). Point clouds were generated using Meta's MultiSLAM. Trajectories are overlaid in orange (head-mounted camera) and green (cane-mounted camera), illustrating differences in environmental coverage and motion patterns between sensor placements.
• Figure 3: NeRF-based reconstructions using three input configurations—head-only, cane-only, and head+cane—across multiple locations. Each row shows one scene: the first three columns depict synthesized views generated from each configuration. The final two columns show the closest ground-truth frames captured by the head-mounted and cane-mounted cameras, respectively, for visual reference. The combined configuration consistently produces sharper, more complete reconstructions, capturing both ground-level geometry and high-frequency environmental details.