• blind
• genetic algorithm
• navigation system
• visual impairment
• wearable device
• wearable haptics

According to the World Health Organization (WHO), around 2.2 billion people in the world are living with vision impairment, of which 43 million are blind. These numbers are destined to increase due to population aging and environmental and lifestyle changes. Blindness has serious consequences on these people's lives, especially in terms of privacy and autonomous navigation; in addition, it depletes social and physical interaction, and moreover, has a significant impact on the family members and carers of these individuals.
Another relevant issue is that vision impairment affects society with a relevant financial burden because of productivity loss.
To overcome these problems, in recent years, several types of devices based on sensory substitution have been developed, which mostly rely on auditory or tactile information. As the major limitation with the types of aid devices that use hearing sense is that they interfere with the auditory stimuli of the environment, touch is more frequently relied on. Portable and wearable tactile devices have then been developed, but since portable devices require to be hand-carried by the user, they are less comfortable. Wearable systems can be more convenient, as they can be easily carried without hindrances and leave users' hands-free. In either case, though, a crucial aspect must be considered, that is the user's real needs and demands, to achieve a system that can be accepted and easily used in real scenarios. The aim of this study is to propose the integration of a high-level (for calculating the global route) and low-level (for performing obstacle avoidance) planning algorithm to be used for indoor environment navigation, which relies on the navigation stimuli delivered with a wearable haptic device and users' requirements.
A genetic algorithm (GA) was developed for the high-level planning task, because very suitable for accomplishing the designated tasks; this algorithm has been integrated with the low-level planning developed in previous research works. This integration has been completed with the usage of a wearable haptic device to deliver navigation cues. For testing the algorithm, I have acquired the map of the second floor of the School of Engineering.
The global paths calculations by the genetic algorithm have been represented and adjusted via simulations on Rviz, then online performances of both the algorithms interacting with each other have been tested preliminarily in the real scenario represented by the map. The results of the simulations validate the effectiveness of the algorithm. Therefore, I tested the whole navigation system, consisting of two cameras, a processing unit, and a haptic device to convey stimuli to the user’s arm, accomplishing preliminary experiments with blindfolded participants. The results presented in this thesis can open promising perspectives for endowing blind and visually impaired people with autonomous navigation capabilities in indoor environments.