Student University of South Florida, United States
Introduction: The prototype aims to serve as a haptic feedback device that replaces hearing with touch, allowing hearing-impaired drivers to navigate safely on the road. The system comprises a sound transducer circuit with RF communication to a portable armband, delivering vibrations as feedback. A complex pattern recognition algorithm forms the core functionality, enabling the device to interpret auditory signals and produce corresponding haptic patterns. This concept could be extended to pedestrians wearing headphones by customizing the algorithm and pattern window to accommodate various periodic auditory signals.
Materials and Methods: The project used Arduino and an imported Arduino FFT library for Fast Fourier Transform on microphone input to detect specific frequencies indicative of an ambulance siren. A sliding window approach with a pattern buffer is used to analyze the frequency data. When a match is found within a tolerance range, the device confirms an ambulance siren and provides haptic and optical feedback via a DC coin motor and LED, transmitted wirelessly using an RF transceiver.
To validate the haptic feedback, we measured the somatosensory neuron response of the distal forearm by using nerve conduction pulse trains. Electrodes are placed on the subject's arm, and baseline measurements are taken. Short electrical impulses and sustained stimulation from the buzzer are used to stimulate the neurons, and their responses are recorded and analyzed.
The overall system consists of a transmitter circuit with a sound sensor connected to Arduino, an RF transceiver, and a receiver circuit with a DC coin motor, LED, and Arduino Nano. The transmitter detects ambulance sirens and wirelessly communicates the alert to the receiver, providing haptic feedback to the user.
Results, Conclusions, and Discussions: To validate the efficacy of the prototypes functionality we tested the detection accuracy under various conditions, including no background noise, room acoustics, and average traffic sounds. The device achieved an average accuracy of 89.44% in detecting the siren pattern, with the room acoustics showing the least accuracy. This is thought to be due to variability in human voices and their natural tendency to have inflections while speaking that may mimic our pattern window frequencies thereby false alerting the device.
To evaluate false detections, we tested the spontaneous firing rate of the algorithm by playing background noises without a siren present. The average false firing rate was recorded at 3.125%, with traffic noise resulting in the highest number of false detections due to frequency similarities.
In the physiological analysis, the team measured the somatosensory neuron response using nerve conduction. The presence of the vibrating motor on the subject's arm led to a rise in membrane potential, causing action potentials to fire, demonstrating the effectiveness of the tactile feedback in stimulating neurons and causing higher cortical processing.
The team overcame challenges in determining the frequency range and tolerance for the siren and optimizing window and buffer sizes for the sliding window approach. The device's portability and affordability make it a potential solution for hearing-impaired drivers to receive emergency signals through haptic feedback.
The device's impact extends beyond assisting hearing-impaired drivers. It could aid in detecting tornado sirens during natural disasters, provide alerts for pedestrians wearing headphones, and improve safety in autonomous vehicles.
Future modifications could include adding more patterns for different warning sounds and integrating the device into vehicles' dashboards or pedestrian headphones and smartwatches.
Overall, the haptic feedback device shows promise in addressing the needs of hearing-impaired individuals and has potential applications in various scenarios. With further development and fine-tuning, it could become a valuable tool for improving safety and communication in different environments.
Acknowledgements (Optional): We would like to thank Dr. Olukemi Akintewe, Zachary Adams, and Randall Buck for all their assistance throughout this project. They were indispensable to the overall success of this team
References (Optional): [1] American Speech-Language-Hearing Association. "Hearing Aids: How Much Do They Cost?" [Online]. Available: https://www.asha.org/public/hearing/hearing-aid-costs/. [Accessed: May 4, 2023].
[2] Dejan, "nRF24L01 - How It Works, Arduino Interface, Circuits, Codes" HowToMechatronics. [Online]. Available: [https://howtomechatronics.com/tutorials/arduino/arduino-wireless-communication-nrf24l01-tutorial/]. [Accessed: May 4, 2023].
[3] Hyundai, "Hyundai Reveals Technology to Assist Hearing-Impaired Drivers," Hyundai News Europe, 2023. [Online]. Available: https://www.hyundai.news/eu/articles/press-releases/hyundai-reveals-technology-to-assist-hearing-impaired-drivers.html. [Accessed: May 4, 2023].
[4] Kosme, 2023, Fast Fourier Transform for Arduino, GitHub, Available: https://github.com/kosme/arduinoFFT . [Accessed: May 4, 2023].
[5] López Montoro, J. s. K. D., Mads Johan (2017). "Automatic Detection of Emergency Cars." Aalborg University Acoustics and Audio technology: 99.
[6] TMRh20, 2023 ,OSI Layer 2 driver for nRF24L01 on Arduino & Raspberry Pi/Linux Devices, GitHub Available: https://github.com/nRF24/RF24 . [Accessed: May 4, 2023].
