Easy Return: An App for Indoor Backtracking Assistance
summary
An iPhone app paired with an Apple Watch records indoor routes as sequences of turn directions and step counts using inertial sensors, a step detector and an HMM-based turn detector trained on the open WeAllWalk dataset of ten blind walkers, and matches return-walk observations to the reversed path with a dynamic‑programming matcher. A controlled study with six blind participants on corridor routes (2–4 turns) shows higher completeness on complex paths; limits include poor open‑space handling and only rudimentary wrong‑turn recovery. Speech feedback and watch controls are used.
AI-written summary · about
chapters (9)
- 0:00Start
- 3:26Step counting
- 4:47Turn detection
- 5:20Path matcher
- 7:19User Interface
- 8:33Experiment
- 9:29Quantitative results
- 10:30Exit survey
- 12:13Conclusions
related talks
27mOne blind weekend
35mHow Android smartphones can help blind people's mobility
58mAccessible Spatial Computing
21mDesigning AR Visualizations to Facilitate Stair Navigation for People with Low Vision
32mDesigning for Accessibility
08mOne Blind Weekend
43mSI@UNIPV: An Inclusive Solution for Autonomous Mobility at the University of Pavia
06mVirtual Paving: Rendering a Smooth Path for People with Visual Impairment through Vibrotactile ...