A navigation system for the visually impaired an intelligent white cane

A Hand-Held Device Presenting Haptic Directional Cues for the Visually Impaired

Haptic information is essential in everyday activities, especially for visually impaired people in terms of real-world navigation. Since human haptic sensory processing is nonlinear, asymmetric vibrations have been widely studied to create a pulling sensation for the delivery of directional haptic cues. However, the design of an input control signal that generates asymmetric vibrations has not yet been parameterised. In particular, it is unclear how to quantify the asymmetry of the output vibrations to create a better pulling sensation. To better understand the design of an input control signal that generates haptic directional cues, we evaluated the effect of the pulling sensations corresponding to the three adjustable parameters (i.e., delay time, ramp-down step length, and cut-off voltage) in a commonly applied step-ramp input signal. The results of a displacement measurement and a psychophysical experiment demonstrate that when the quantified asymmetry ratio is in a range of 0.3430-0.3508 with an optimised cut-off voltage for our hand-held device, the haptic directional cues are better perceived by participants. Additionally, the results also showed a superior performance in haptic delivery by shear forces than normal forces.

Efficient Multi-Object Detection and Smart Navigation Using Artificial Intelligence for Visually Impaired People

Visually impaired people face numerous difficulties in their daily life, and technological interventions may assist them to meet these challenges. This paper proposes an artificial intelligence-based fully automatic assistive technology to recognize different objects, and auditory inputs are provided to the user in real time, which gives better understanding to the visually impaired person about their surroundings. A deep-learning model is trained with multiple images of objects that are highly relevant to the visually impaired person. Training images are augmented and manually annotated to bring more robustness to the trained model. In addition to computer vision-based techniques for object recognition, a distance-measuring sensor is integrated to make the device more comprehensive by recognizing obstacles while navigating from one place to another. The auditory information that is conveyed to the user after scene segmentation and obstacle identification is optimized to obtain more information in less time for faster processing of video frames. The average accuracy of this proposed method is 95.19% and 99.69% for object detection and recognition, respectively. The time complexity is low, allowing a user to perceive the surrounding scene in real time.

Egocentric Landmark-Based Indoor Guidance System for the Visually Impaired

In this paper, we introduce an egocentric landmark-based guidance system that enables visually impaired users to interact with indoor environments. The user who wears Google Glasses will capture his surroundings within his field of view. Using this information, we provide the user an accurate landmark-based description of the environment including his relative distance and orientation to each landmark. To achieve this functionality, we developed a near real time accurate vision based localization algorithm. Since the users are visually impaired our algorithm accounts for captured images using Google Glasses that have severe blurriness, motion blurriness, low illumination intensity and crowd obstruction. We tested the algorithm performance in a 12,000 ft2 open indoor environment. When we have mint query images our algorithm obtains mean location accuracy within 5ft., mean orientation accuracy less than 2 degrees and reliability above 88%. After applying deformation effects to the query images such blurriness, motion blurriness and illumination changes, we observe that the reliability is above 75%.

Real-time dangling objects sensing: A preliminary design of mobile headset ancillary device for visual impaired

Blinds and severe visual impairments can utilize tactile sticks to assist their walking. However, they cannot fully understand the dangling objects in front of their walking routes. This research proposed a mobile real-time dangling objects sensing (RDOS) prototype, which is located on the cap to sense any front barrier. This device utilized cheap ultrasonic sensor to act as another complement eye for blinds to understand the front dangling objects. Meanwhile, the RDOS device can dynamically adjust the sensor's front angle that is depended on the user's body height and promote the sensing accuracy. Meanwhile, two major required algorithms, height-angle measurement and ultrasonic sensor alignment, are proposed with this prototype. The research team also integrated the RDOS device prototype with mobile Android devices by communicating with Bluetooth to record the walking route.