Given the growth in the numbers of visually impaired (VI) people in low-income countries, the development of affordable electronic travel aid (ETA) systems employing devices, sensors, and apps embedded in ordinary smartphones becomes a potentially cost-effective and reasonable all-in-one solution of utmost importance for the VI. This paper offers an overview of recent ETA research prototypes that employ smartphones for assisted orientation and navigation in indoor and outdoor spaces by providing additional information about the surrounding objects. Scientific achievements in the field were systematically reviewed using PRISMA methodology. Comparative meta-analysis showed how various smartphone-based ETA prototypes could assist with better orientation, navigation, and wayfinding in indoor and outdoor environments. The analysis found limited interest among researchers in combining haptic interfaces and computer vision capabilities in smartphone-based ETAs for the blind, few attempts to employ novel state-of-the-art computer vision methods based on deep neural networks, and no evaluations of existing off-the-shelf navigation solutions. These results were contrasted with findings from a survey of blind expert users on their problems in navigating in indoor and outdoor environments. This revealed a major mismatch between user needs and academic development in the field.

Blind persons need electronic traveling aid (ETA) solutions for better orientation and navigation in unfamiliar indoor environments, with embedded features for detection and recognition of both obstacles and desired destinations such as rooms, staircases, and elevators. Because the use of GPS for locational references is impractical indoors, the development of such navigation systems is challenging and requires a systematic review and evaluation of different technological approaches. Using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method, we evaluated and compared current research papers that deal with the prototyping of assistive devices (visual sensory perception substitution with audio and haptic signals) for blind and visually impaired persons. We conducted an instructional assessment of selected indoor navigation prototypes using three main criteria: navigation technologies, sensors, and computer vision approaches. For the latter category, we conducted a separate systematic review, as papers in this research area primarily specialize in software computer vision solutions rather than hardware. The paper provides useful insights for researchers regarding technological instrumentation for the development of ETA solutions for blind and visually impaired (VI) persons in the field of indoor orientation and navigation.

End-to-end autonomous driving often relies on the concept of learning to imitate from expert demonstrations, which is imitation learning. Since expert demonstrations cannot be infinite and cover all possible variations in data, there always are situations where the trained agents encounter unseen conditions. The Encountering of unseen conditions causes a shift in the data distribution. One of the most common causes of this shift is changes in weather and lighting conditions. We point out that the driving agent can be prepared for such variations in data distribution without solely relying on data aggregation methods, which has been the solution so far. We propose a method which introduces a new step, i.e., pre-training of the imitation learning-based driving agent over the task of visual place recognition (VPR), which incorporates exposure to place images under varying weather conditions. To demonstrate our proposed VPR pre-trained method, we train an agent with the proposed pre-training and compare it to a baseline agent which relies on the commonly used ImageNet pre-training method. We also evaluate the trained agents as per the Leaderboard driving benchmark in CARLA environment. Our results show, pre-training on the VPR task shows higher resistance to unseen weather conditions. The findings calculated by us are evaluated over multiple seeds to show statistical consistency. The accompanying open source code repository can by accessed via https://github.com/Shubhamcl/vpr_pretrained_agent/.

Publisher: Vilnius University Institute of Data Science and Digital Technologies, Journal: Informatica, Title: Neural Network with Matrix Inputs, Authors: Povilas Daniušis, Pranas Vaitkus , In this paper we propose and analyze a multilayer perceptron-like model with matrix inputs. We applied the proposed model to the financial time series prediction problem, compared it with the standard multilayer perceptron model, and got fairly good results.