Main content

Home

Menu

Loading wiki pages...

View
Wiki Version:
# **Low-Cost Open Source Ultrasound-Sensing Based Navigational Support for the Visually Impaired** > *Aliaksei L. Petsiuk and Joshua M. Pearce* <br /> *Michigan Technological > University* <br />*March 2019* --- # **Abstract** Nineteen million Americans have significant vision loss. Over 70% of these are not employed full-time, and more than a quarter live below the poverty line. Globally, there are 36 million blind people, but less than half use white canes or more costly commercial sensory substitutions. The quality of life for visually impaired people is hampered by the resultant lack of independence. To help alleviate these challenges this study reports on the development of a low-cost, open-source ultrasound-based navigational support system in the form of a wearable bracelet to allow people with the lost vision to navigate, orient themselves in their surroundings and avoid obstacles when moving. The system can be largely made with digitally distributed manufacturing using low-cost 3-D printing/milling. It conveys point-distance information by utilizing the natural active sensing approach and modulates measurements into haptic feedback with various vibration patterns within the four-meter range. It does not require complex calibrations and training, consists of the small number of available and inexpensive components, and can be used as an independent addition to traditional tools. Sighted blindfolded participants successfully demonstrated the device for nine primary everyday navigation and guidance tasks including indoor and outdoor navigation and avoiding collisions with other pedestrians. <br /><br /> # **Design** The system conveys point-distance information by utilizing the natural active sensing approach and modulates measurements into haptic feedback with various vibration patterns within the distance range of 3 meters. @[osf](48rsq) **Figure 1.** Parts of an open-source navigational support with 3-D printable case components: (a) 3D prototype; (b) Assembly; (c) Model 1 with one vibration motor; (d) Model 2 with two vibration motors; (e) Locking rings; (f) Case; (g) Vibration pad; (h) Sensor core; (i) Back cap; (j) Bracelet. @[osf](jc4d6) **Figure 2.** Assembly of an open-source navigational support with 3-D printable case components. @[osf](52qam) **Figure 3.** Devices in use. The hand bracelet (Figure 1) has an [online option for customization][1], so a person with no experience with complicated 3-D modeling software could print the part after adjusting it to their hand size. <br /><br /> # **Operating principle** Assembly instructions are available at the [Appropedia page][2]. @[osf](4f9bj) **Figure 4.** The ultrasonic sensor operating principles: (a) The principal distances (not to scale); (b) Calibration of the optimal duty cycle equation for the distance range of 35 cm to 150 cm, where (c) MDC = 127 + 127 · tanh (-(D - 70) / 35); (d) MDC = 127 + 127 · tanh (-(D - 150) / 35); (e) MDC = 296 – 1.5 · D; (f) MDC = 335 – 1.3 · D; (g) MDC = -77 + 2.2 · D; (h) MDC = -48 + 1.2 · D. The measured distance is modulated with vibration amplitude and translated in real-time as a duty cycle parameter from the Arduino board (Figure 4, a). Distances up to 35 cm are characterized by single vibration pulses with a relatively high frequency. Distances from 150 to 250 cm are characterized by single pulses with low periodicity, and distances above 250 cm are modulated with two-pulse beats (Figure 4, b). An optimal duty cycle equation for the most common distance range of 35-150 cm was found during experiments and calibrations (Figure 5). The generated duty cycle for the Arduino output, MDC is: *MDC = m + m · tanh (-(D - k) / b) = 127 + 127 · tanh (-(D - 70) / 35) = 0...255,* Where m = 127, k = 70 and b = 35 are the calibrated parameters, and D is the measured distance in the range of 35 cm to 150 cm. @[osf](7cuhq) **Figure 5.** Calibrating procedure: (a) Hand swinging; (b) Wall following; (c) Obstacle detection; (d) Curbs tracking. <br /><br /> # **Electronics and firmware** @[osf](w9bxs) **Figure 6.** Electrical circuit. The Arduino sketch for the sensor core (Figure 1), `usound_navigational_support__tanh_loop.ino`, is in this repository. <br /><br /> # **Testing and experimental results** @[youtube](https://www.youtube.com/watch?v=FA9r2Y27qvY) <br /><br /> # **Conclusions** The developed low-cost (< 24 USD), open-source navigational support system allows people with lost vision to solve the primary tasks of navigation, orientation, and obstacle detection (>0.5 m2 stationary within the distance range of up to 4 m and moving up to 0.5 m/s within the distance range of up to 1 m) to ensure their safety and mobility. The devices demonstrated intuitive haptic feedback, which becomes easier to use with short practice. It can be largely digitally manufactured as an independent device or as a complementary part to the available means of sensory augmentation (e.g., a white cane). The device operates in similar distance ranges as most of the observed commercial products, and it can be replicated by a person without high technical qualification. Since the prices for available commercial products vary from $90–800 USD, the cost savings ranged from a minimum of 73.5% to over 97%. [1]: https://www.thingiverse.com/thing:3733136 [2]: https://www.appropedia.org/Low-cost_open_source_ultrasound-sensing_based_navigational_support_for_visually_impaired