Ultrasonic Obstacle Detection Device!
Ultrasonic Obstacle Detection Device!
Here is our paper on our device for information and our results in testing:
Our 2 testing methodologies:
Directional sensing test as we have 2 sensors and 1 vibration motor on each side to distinguish direction
Setup: Blind participant holds the dual-sonar device (HC-SR04 sensors ~30 cm apart) at chest height, facing a blank wall. A sighted “walker” stands hidden behind a barrier (e.g. a low screen) at a fixed distance (≈4 m), directly aligned with one of the tape lines on the floor. Two parallel tape lines—one extending straight from each sensor—mark the paths the walker will follow toward the participant’s left or right sensor.
Procedure: On each trial, the walker chooses one of the two tape lines, walks straight toward the device-holder at a moderate, constant pace (≈3 mph), and stops at ≈0.5 m. The blind participant:
Calls out the moment they first feel any vibration.
Rates the intensity: e.g. on a 1–5 scale, every meter of approach.
States direction: “coming from my left” or “coming from my right.”
Data Collected:
Actual distance at first vibration (logged via microcontroller).
Vibration PWM vs. distance curve (for both left and right motors).
Participant’s directional accuracy (% correct).
Self-Guided Wall Approach (Calibration & Distance Mapping)
Setup: Blind participant alone, device mounted on a chest harness, stands on a single tape line facing a flat wall at ≈4 m.
Procedure: Participant walks themself straight toward the wall at a comfortable pace (≈2–3 mph). They stop when they feel the strongest sustained vibration or when they decide. Repeat from starting distances of 4 m, 3 m, and 2 m (5 trials each, order randomized).
Participant Task: For each trial, physically note:
Distance of first vibration (“ participant did thumbs up at about X meters”).
Intensity profile: how vibration felt as they closed the final meters (rated 1–5).
Confidence (1–5) that the wall was what triggered the vibration.
Data Collected:
True distance at vibration onset (logged by sensor).
Vibration PWM vs. distance calibration curve.
Subjective distance estimates vs. actual.