基于单片机与深度学习的人数统计与密度预警系统

在智慧城市与智能安防快速发展的背景下，多摄像头人物追踪技术成为提升公共安全管理效率的核心需求。传统方案存在跨摄像头身份关联失效、计算资源消耗大及实时性不足等问题。本文提出一种基于边缘计算与中心数据库协同架构的实时人物追踪系统，采用轻量化模型YOLOv5s进行人体检测，结合OSNet_x0.25提取512维特征向量，通过向量数据库实现毫秒级特征匹配。实验表明，该系统在Market-1501数据集上Rank-1准确率达86.7%，640 × 640视频流实时处理帧率为8.2 FPS，在校园场景中验证了工程实用性。

本文聚焦于基于多目标跟踪的课堂人数自动统计这一主题，提出将YOLOv8和SORT (Simple Online and Realtime Tracking)算法相结合的方法。其中，YOLOv8作为先进的单阶段目标检测算法，能迅速识别图像或视频帧中的目标并输出关键信息；SORT算法借助卡尔曼滤波预测目标位置变化。研究通过运用特定的软硬件配置开展实验，对模型进行训练与验证。该方法旨在实现课堂人数的精准自动统计，满足课堂管理和动态监控的需求，为教师及管理人员提供有力工具。其不仅在算法应用上具有创新性，而且在教育管理实践中具有重要的实用价值，有望对智慧教育领域的课堂管理产生积极深远的影响。

随着社会的不断发展，人群密集的场所随处可见。对监控视频下的人员进行统计分析，实现人数统计算法，可以为城市公共资源优化配置、安保人员调度、安全管理等提供有效的技术手段。本文基于YOLO V4平台，采用深度学习算法来识别监控视频的人，并加入统计算法对识别出的人进行统计计算，实现视频监控下的人数统计。

随着城市交通管理、安保和智慧城市建设的需求日益增长，研究高效、准确、实时的行人定位和追踪系统具有重要的研究意义和实用价值。本文针对目前计算机视觉和人工智能领域的行人检测和跟踪技术，提出了一个基于视觉的行人定位和追踪系统。系统采用YOLOv5算法进行行人检测，DeepSORT算法进行多目标跟踪，并使用ResNet50模型实现行人重识别。此外，系统还支持目标运动轨迹绘制、行人数量统计等功能。实验结果表明，该系统在公开数据集(例如Market-1501数据集)和现实场景数据上具有良好的性能表现，可为城市交通管理、安保和智慧城市建设等领域提供有力支持。未来，将继续优化算法和模型结构，提高系统的准确性和实时性，为计算机视觉和人工智能领域的研究和应用带来更多创新。

在机器人辅助老年人淋浴的环境中，水雾的存在会掩盖关键身体点的视线，对精确人体关键点检测构成挑战。本研究提出了一种端到端模型，将人类关键点检测与去雾技术相结合。该模型旨在监测老年人在淋浴过程中的关键点，从而防止事故发生。为了评估我们方法的有效性，我们制作了一个具有细水雾的模拟数据集，并比较了各种模型。实验结果表明，与我们精选数据集上的其他模型相比，我们提出的模型实现了82.63%的平均精度均值(mAP)的显著提高，比基线提高了23%。此外，它还可以有效监测老年人的淋浴过程。这项研究具有实际意义，并有可能提高老年人的生活质量。

Ultralow-resolution infrared (IR) array sensors offer a low cost, energy efficient, and privacy-preserving solution for people counting, with applications, such as occupancy monitoring and visitor flow analysis in private and public spaces. Previous work has shown that deep learning (DL) can yield superior performance on this task. However, the literature was missing an extensive comparative analysis of various efficient DL architectures for IR array-based people counting, that considers not only their accuracy but also the cost of deploying them on memory- and energy-constrained Internet of Things (IoT) edge nodes. Such analysis is key for system designers, since it helps them select the most appropriate DL model given the constraints of their target hardware. In this work, we address this need by comparing six different DL architectures on a novel data set composed of IR images collected from a commercial <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$8\times8$ </tex-math></inline-formula> array, which we made openly available. With a wide architectural exploration of each model type, we obtain a rich set of Pareto-optimal solutions, spanning cross-validated balanced accuracy scores in the 55.70%–82.70% range. When deployed on a commercial microcontroller (MCU) by STMicroelectronics, the STM32L4A6ZG, these models occupy 0.41–9.28kB of memory, and require 1.10–7.74 ms per inference, while consuming 17.18– <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$120.43 \mu \text{J}$ </tex-math></inline-formula> of energy. Our models are significantly more accurate than a previous deterministic method (up to +39.9%), while being up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$3.53\times $ </tex-math></inline-formula> faster and more energy efficient. So, our work serves also as a demonstration that DL can not only achieve higher accuracy but also higher efficiency compared to classic algorithms for this type of task. Further, our models’ accuracy is comparable to state-of-the-art DL solutions on similar resolution sensors, despite a much lower complexity. All our models enable continuous, real-time inference on an MCU-based IoT node, with years of autonomous operation without battery recharging.

Recent spectacular progress in computational technologies has led to an unprecedented boom in the field of Artificial Intelligence (AI). AI is now used in a plethora of research areas and has demonstrated its capability to bring new approaches and solutions to various research problems. However, the extensive computation required to train AI algorithms comes with a cost. Driven by the need to reduce the energy consumption, the carbon footprint and the cost of computers running machine learning algorithms, TinyML is nowadays considered as a promising AI alternative focusing on technologies and applications for extremely low-profile devices. This paper presents the results of a literature survey of all TinyML applications and related research efforts. Our survey builds a taxonomy of TinyML techniques that have been used so far to bring new solutions to various domains, such as healthcare, smart farming, environment, and anomaly detection. Finally, this survey highlights the remaining challenges and points out possible future research directions. We anticipate that this survey will motivate further discussions on the various fields of applications of TinyML and the synergy of resource-constrained devices and edge intelligence.

In a few years, the world will be populated by billions of connected devices that will be placed in our homes, cities, vehicles, and industries. Devices with limited resources will interact with the surrounding environment and users. Many of these devices will be based on machine learning models to decode meaning and behavior behind sensors' data, to implement accurate predictions and make decisions. The bottleneck will be the high level of connected things that could congest the network. Hence, the need to incorporate intelligence on end devices using machine learning algorithms. Deploying machine learning on such edge devices improves the network congestion by allowing computations to be performed close to the data sources. The aim of this work is to provide a review of the main techniques that guarantee the execution of machine learning models on hardware with low performances in the Internet of Things paradigm, paving the way to the Internet of Conscious Things. In this work, a detailed review on models, architecture, and requirements on solutions that implement edge machine learning on Internet of Things devices is presented, with the main goal to define the state of the art and envisioning development requirements. Furthermore, an example of edge machine learning implementation on a microcontroller will be provided, commonly regarded as the machine learning "Hello World".

With the growing number of deployments of Internet of Things (IoT) infrastructure for a wide variety of applications, the battery maintenance has become a major limitation for the sustainability of such infrastructure. To overcome this problem, energy harvesting offers a viable alternative to autonomously power IoT devices, resulting in a number of battery-less energy harvesting IoTs (or EH-IoTs) appearing in the market in recent years. Standards activities are also underway, which involve wireless protocol design suitable for EH-IoTs as well as testing procedures for various energy harvesting methods. Despite the early commercial and standards activities, IoT sensing, computing and communications under unpredictable power supply still face significant research challenges. This paper systematically surveys recent advances in EH-IoTs from several perspectives. First, it reviews the recent commercial developments for EH-IoT in terms of both products and services, followed by initial standards activities in this space. Then it surveys methods that enable the use of energy harvesting hardware as a proxy for conventional sensors to detect contexts in energy efficient manner. Next it reviews the advancements in efficient checkpointing and timekeeping for intermittently powered IoT devices. We also survey recent research in novel wireless communication techniques for EH-IoTs, such as the applications of reinforcement learning to optimize power allocations on-the-fly under unpredictable energy productions, and packet-less IoT communications and backscatter communication techniques for energy impoverished environments. The paper is concluded with a discussion of future research directions.

No abstract

No abstract

For your convenience Apress has placed some of the front matter material after the index.

This review paper discusses the trends and projections for wearable technology in the consumer sports sector (excluding professional sport). Analyzing the role of wearable technology for different users and why there is such a need for these devices in everyday lives. It shows how different sensors are influential in delivering a variety of readings that are useful in many ways regarding sport attributes. Wearables are increasing in function, and through integrating technology, users are gathering more data about themselves. The amount of wearable technology available is broad, each having its own role to play in different industries. Inertial measuring unit (IMU) and Global Positioning System (GPS) sensors are predominantly present in sport wearables but can be programmed for different needs. In this review, the differences are displayed to show which sensors are compatible and which ones can evolve sensor technology for sport applications.

Recognizing people by their gait has become more and more popular nowadays due to the following reasons. First, gait recognition can work well remotely. Second, gait recognition can be done from low-resolution videos and with simple instrumentation. Third, gait recognition can be done without the cooperation of individuals. Fourth, gait recognition can work well while other features such as faces and fingerprints are hidden. Finally, gait features are typically difficult to be impersonated. Recent ubiquity of smartphones that capture gait patterns through accelerometers and gyroscope and advances in machine learning have opened new research directions and applications in gait recognition. A timely survey that addresses current advances is missing. In this article, we survey research works in gait recognition. In addition to recognition based on video, we address new modalities, such as recognition based on floor sensors, radars, and accelerometers; new approaches that include machine learning methods; and examine challenges and vulnerabilities in this field. In addition, we propose a set of future research directions. Our review reveals the current state-of-art and can be helpful to both experts and newcomers of gait recognition. Moreover, it lists future works and publicly available databases in gait recognition for researchers.

Alzheimer's disease (AD) stands as the predominant form of dementia, presenting significant and escalating global challenges. Its etiology is intricate and diverse, stemming from a combination of factors such as aging, genetics, and environment. Our current understanding of AD pathologies involves various hypotheses, such as the cholinergic, amyloid, tau protein, inflammatory, oxidative stress, metal ion, glutamate excitotoxicity, microbiota-gut-brain axis, and abnormal autophagy. Nonetheless, unraveling the interplay among these pathological aspects and pinpointing the primary initiators of AD require further elucidation and validation. In the past decades, most clinical drugs have been discontinued due to limited effectiveness or adverse effects. Presently, available drugs primarily offer symptomatic relief and often accompanied by undesirable side effects. However, recent approvals of aducanumab (1) and lecanemab (2) by the Food and Drug Administration (FDA) present the potential in disrease-modifying effects. Nevertheless, the long-term efficacy and safety of these drugs need further validation. Consequently, the quest for safer and more effective AD drugs persists as a formidable and pressing task. This review discusses the current understanding of AD pathogenesis, advances in diagnostic biomarkers, the latest updates of clinical trials, and emerging technologies for AD drug development. We highlight recent progress in the discovery of selective inhibitors, dual-target inhibitors, allosteric modulators, covalent inhibitors, proteolysis-targeting chimeras (PROTACs), and protein-protein interaction (PPI) modulators. Our goal is to provide insights into the prospective development and clinical application of novel AD drugs.

Growth of population and the inception of new devices every day comes with an incessant rise in energy consumption and has brought great challenges in terms of energy management at the consumer side. With the evolution of technology, smart meters (SMs) are not only considered merely as tools to measure energy consumption but act as a main resource of energy management systems. The application of SM spans over a wide range of advantages, including accurate billing data, information of utilization at the user end, the establishment of two-way communication and remote control of the user equipment. SM is the most essential element of a smart power grid that with the help of any smart energy management system (SEMS), assesses, measures, controls, implements and communicates power allocation, utilization, and consumption at both, single device, and network level. Data provided by the SMs is used by power supply companies to revolutionize power distribution and consumption through various techniques such as, non-intrusive load monitoring and demand-side management (DSM). For efficient data gathering and utilization, internet of things (IoT) is emerging as a key partner in the power industry leading to effective resource management. This paper provides presentation, deployment, and validation of an IoT based SEMS strategy and its related benefits to overcome challenges of energy management at consumer side. The presented SEMS incorporates various communication interfaces and protocols to integrate with any software-based smart solution. In this work, Entrack software is used for data gathering and analysis. As a proof of concept, the presented system is tested and implemented in 4 different buildings of a well-known private company in Pakistan, i.e., Stylo Pvt. Ltd. The case study analysis in this work shows the effectiveness of the presented IoT based SEMS.

No abstract

The increasing food scarcity necessitates sustainable agriculture achieved through automation to meet the growing demand. Integrating the Internet of Things (IoT) and Wireless Sensor Networks (WSNs) is crucial in enhancing food production across various agricultural domains, encompassing irrigation, soil moisture monitoring, fertilizer optimization and control, early-stage pest and crop disease management, and energy conservation. Wireless application protocols such as ZigBee, WiFi, SigFox, and LoRaWAN are commonly employed to collect real-time data for monitoring purposes. Embracing advanced technology is imperative to ensure efficient annual production. Therefore, this study emphasizes a comprehensive, future-oriented approach, delving into IoT-WSNs, wireless network protocols, and their applications in agriculture since 2019. It thoroughly discusses the overview of IoT and WSNs, encompassing their architectures and summarization of network protocols. Furthermore, the study addresses recent issues and challenges related to IoT-WSNs and proposes mitigation strategies. It provides clear recommendations for the future, emphasizing the integration of advanced technology aiming to contribute to the future development of smart agriculture systems.

Security of lives and properties is highly important for enhanced quality living. Smart home automation and its application have received much progress towards convenience, comfort, safety, and home security. With the advances in technology and the Internet of Things (IoT), the home environment has witnessed an improved remote control of appliances, monitoring, and home security over the internet. Several home automation systems have been developed to monitor movements in the home and report to the user. Existing home automation systems detect motion and have surveillance for home security. However, the logical aspect of averting unnecessary or fake notifications is still a major area of challenge. Intelligent response and monitoring make smart home automation efficient. This work presents an intelligent home automation system for controlling home appliances, monitoring environmental factors, and detecting movement in the home and its surroundings. A deep learning model is proposed for motion recognition and classification based on the detected movement patterns. Using a deep learning model, an algorithm is developed to enhance the smart home automation system for intruder detection and forestall the occurrence of false alarms. A human detected by the surveillance camera is classified as an intruder or home occupant based on his walking pattern. The proposed method’s prototype was implemented using an ESP32 camera for surveillance, a PIR motion sensor, an ESP8266 development board, a 5 V four‐channel relay module, and a DHT11 temperature and humidity sensor. The environmental conditions measured were evaluated using a mathematical model for the response time to effectively show the accuracy of the DHT sensor for weather monitoring and future prediction. An experimental analysis of human motion patterns was performed using the CNN model to evaluate the classification for the detection of humans. The CNN classification model gave an accuracy of 99.8%.

Computing has undergone a significant transformation over the past two decades, shifting from a machine-based approach to a human-centric, virtually invisible service known as ubiquitous or pervasive computing. This change has been achieved by incorporating small embedded devices into a larger computational system, connected through networking and referred to as edge devices. When these devices are also connected to the Internet, they are generally named Internet-of-Thing (IoT) devices. Developing Machine Learning (ML) algorithms on these types of devices allows them to provide Artificial Intelligence (AI) inference functions such as computer vision, pattern recognition, etc. However, this capability is severely limited by the device's resource scarcity. Embedded devices have limited computational and power resources available while they must maintain a high degree of autonomy. While there are several published studies that address the computational weakness of these small systems-mostly through optimization and compression of neural networks- they often neglect the power consumption and efficiency implications of these techniques. This study presents power efficiency experimental results from the application of well-known and proven optimization methods using a set of well-known ML models. The results are presented in a meaningful manner considering the "real world" functionality of devices and the provided results are compared with the basic "idle" power consumption of each of the selected systems. Two different systems with completely different architectures and capabilities were used providing us with results that led to interesting conclusions related to the power efficiency of each architecture.

Mosques are worship places of Allah and must be preserved clean, immaculate,\nprovide all the comforts of the worshippers in them. The prophet's mosque in\nMedina/ Saudi Arabia is one of the most important mosques for Muslims. It\noccupies second place after the sacred mosque in Mecca/ Saudi Arabia, which is\nin constant overcrowding by all Muslims to visit the prophet Mohammad's tomb.\nThis paper aims to propose a smart dome model to preserve the fresh air and\nallow the sunlight to enter the mosque using artificial intelligence\ntechniques. The proposed model controls domes movements based on the weather\nconditions and the overcrowding rates in the mosque. The data have been\ncollected from two different resources, the first one from the database of\nSaudi Arabia weather's history, and the other from Shanghai Technology\nDatabase. Congested Scene Recognition Network (CSRNet) and Fuzzy techniques\nhave applied using Python programming language to control the domes to be\nopened and closed for a specific time to renew the air inside the mosque. Also,\nthis model consists of several parts that are connected for controlling the\nmechanism of opening/closing domes according to weather data and the situation\nof crowding in the mosque. Finally, the main goal of this paper has been\nachieved, and the proposed model has worked efficiently and specifies the exact\nduration time to keep the domes open automatically for a few minutes for each\nhour head.\n

Integrating embedded cyber-physical systems (CPS) in smart energy and utilities provides extensive security. Most applications in smart cities require supercomputing powers for model performance. Security applications like crowd monitoring and threat detection require cyber-physical integration with camera operations. Recording and analysing the data in real time consumes enormous power and energy. This enables AI-powered crowd-tracking and threat-detection systems with real-time capabilities and reduced energy usage. Though several Literature studies are presented on energy-aware applications, this paper exhibits a novel system wherein embedded CPS, deep learning, and smart energy systems are integrated to solve public safety systems for conserving energy. Localised employees collect real-time data points by operating IoT sensors, cameras, and microcontroller-powered CPS nodes, which use edge computing and fog computing to reduce latency. Edge computing collects and processes sensitive data locally by minimising external network exposure. Fog computing strengthens security by decentralising data processing. Crowd density estimation, behavior prediction, and anomaly detection are performed using Convolutional Neural Networks (CNNs). A smart munitions trace system is proposed, which utilizes energy from smart grids and renewable sources to power distributed CPS nodes, applies fast response to allocate resources under crowd movements, and integrates with smart grids. Threat detection monitoring and management algorithms detect suspect activities or objects and issue appropriate notifications quickly to facilitate effective emergency response. Along with these, cloud-assisted deep learning models provide infrastructure development over time. This multi-disciplinary method solves issues like data privacy using local edge computing. This study enhances innovative city initiatives by integrating embedded CPS and energy technologies for safety measures with energy management.

The advent of machine learning (ML) methods for the industry has opened new possibilities in the automotive domain, especially for Advanced Driver Assistance Systems (ADAS). These methods mainly focus on specific problems ranging from traffic sign and light recognition to pedestrian detection. In most cases, the computational resources and power budget found in ADAS systems are constrained while most machine learning methods are computationally intensive. The usual solution consists in adapting the ML models to comply with the memory and real-time (RT) requirements for inference. Some models are easily adapted to resource-constrained hardware, such as Support Vector Machines, while others, like Neural Networks, need more complex processes to fit into the desired hardware. The ADAS hardware (HW platforms) are diverse, from complex MPSoC CPUs down to classical MCUs, DPSs and application-specific FPGAs and ASICs or specific GPU platforms (such as the NVIDIA families Tegra or Jetson). Therefore, there is a tradeoff between the complexity of the ML model implemented and the selected platform that impacts the performance metrics: function results, energy consumption and speed (latency and throughput). In this paper, a survey in the form of systematic review is conducted to analyze the scope of the published research works that embed ML models into resource-constrained implementations for ADAS applications and what are the achievements regarding the ML performance, energy and speed trade-off.

Recent years have witnessed the widespread popularity of Internet of things (IoT). By providing sufficient data for model training and inference, IoT has promoted the development of artificial intelligence (AI) to a great extent. Under this background and trend, the traditional cloud computing model may nevertheless encounter many problems in independently tackling the massive data generated by IoT and meeting corresponding practical needs. In response, a new computing model called edge computing (EC) has drawn extensive attention from both industry and academia. With the continuous deepening of the research on EC, however, scholars have found that traditional (non-AI) methods have their limitations in enhancing the performance of EC. Seeing the successful application of AI in various fields, EC researchers start to set their sights on AI, especially from a perspective of machine learning, a branch of AI that has gained increased popularity in the past decades. In this article, we first explain the formal definition of EC and the reasons why EC has become a favorable computing model. Then, we discuss the problems of interest in EC. We summarize the traditional solutions and hightlight their limitations. By explaining the research results of using AI to optimize EC and applying AI to other fields under the EC architecture, this article can serve as a guide to explore new research ideas in these two aspects while enjoying the mutually beneficial relationship between AI and EC.

Efficient real-time computer vision-based passenger flow analysis is increasingly important for the management of intelligent transportation systems and smart cities. This paper presents the design and implementation of a system for real-time object detection, tracking, and people counting in tram stations. The proposed approach integrates YOLO-based detection with a lightweight tracking module and is deployed on an NVIDIA Jetson Nano device, enabling operation under resource constraints and demonstrating the potential of edge AI. Multiple YOLO versions, from v3 to v11, were evaluated on data collected in collaboration with Metropolitano de Tenerife. Experimental results show that YOLOv5s achieves the best balance between detection accuracy and inference speed, reaching 96.85% accuracy in counting tasks. The system demonstrates the feasibility of applying edge AI to monitor passenger flow in real time, contributing to intelligent transportation and smart city initiatives.

In a world where resources are scarce and urban areas consume the vast majority of these resources, it is vital to make cities greener and more sustainable. Advanced systems to improve and automate processes within a city will play a leading role in smart cities. From smart design of buildings, which capture rain water for later use, to intelligent control systems, which can monitor infrastructures autonomously, the possible improvements enabled by sensing technologies are immense. Ubiquitous sensing poses numerous challenges, which are of a technological or social nature. This paper presents an overview of the state of the art with regards to sensing in smart cities. Topics include sensing applications in smart cities, sensing platforms and technical challenges associated with these technologies. In an effort to provide a holistic view of how sensing technologies play a role in smart cities, a range of applications and technical challenges associated with these applications are discussed. As some of these applications and technologies belong to different disciplines, the material presented in this paper attempts to bridge these to provide a broad overview, which can be of help to researchers and developers in understanding how advanced sensing can play a role in smart cities.

The coronavirus disease 19 (COVID-19) pandemic has created cataclysmic repercussions in virtually every facet of life and has had profound effects on the practice of medicine. This is particularly true for providers who treat disability. Although all disabilities are unquestionably challenged in undue ways, this perspective is meant to draw special attention to those with visual impairment. COVID-19 is extremely contagious and has spread globally with unprecedented rapidity. The best current countermeasures include personal protective equipment (PPE), social distancing, and minimizing or avoiding touch or contact with surfaces and/or objects that may be contaminated with viral particles, all of which pose unique challenges for those with low or no vision. Co-authors Rizzo and Giudice, themselves visually impaired, are researchers who are investigating creative innovation to combat the untoward consequences of visual impairment, However, this situation transcends their professional interests, as it has directly affected their lives and the lives of other blind individuals close to them. This essay builds on the combination of their personal experiences and research expertise to motivate the current problem and pose some viable solutions. When you cannot see what is around you, touch becomes the primary mode of both exploring and interacting with the environment. We rely on touch to support many tasks throughout the day, whether it be the movement of the keys as we type on our computer, the warmth and heft we feel as we pick up our morning mug of coffee, or the texture of our clothes. However, for blind and visually impaired (BVI) people, the sense of touch and use of haptics (ie, information that is perceived through active touch) transcends these “normal” uses of this modality. For this community, touch perception supports many of the same tasks that sighted people perform on the basis of visual perception. Although hearing and touch represent the principal modes of nonvisual sensing, touch and vision share the ability to accurately convey spatial information. Despite touch having a much smaller “tactile field of view” and lower sensory bandwidth capacity than vision, a growing body of evidence suggests that spatial information learned from both modalities develops into an amodal “spatial image” in the brain that functions equivalently in the service of action, irrespective of the input source.1 This functional equivalence (ie, statistically indistinguishable performance) between touch and vision has been demonstrated for a broad range of spatial behaviors.2 Neuroscientific evidence also corroborates this notion because the same expert processing region of the brain, the parahippocampal place area (PPA), has been found to be preferentially involved during functional magnetic resonance imaging (fMRI) in the computation of spatial layouts learned through haptic and visual perception.3 This study also found no difference in the pattern of neural activation between blind and sighted participants on the haptic tasks, which agrees with other neuroimaging research studying haptic spatial processing in “expert” brain regions between blind and sighted participants.3-5 In aggregate, the evidence showing similarity of behavior after haptic and visual learning and common neural networks underlying spatial computations between blind and sighted individuals provides converging support for the similarity of these senses in the encoding and processing of spatial information, irrespective of visual experience.3-5 One may think of spatial information as the “common denominator” of the senses, with haptic and visual inputs informing us about a common physical space (ie, our perception of the surrounding world). An important consequence of this sensory similarity is that BVI people rely far more heavily on their sense of touch to support spatial tasks than their sighted peers. These tasks may be small-scale, for example, exploring what is on a table, or large-scale, for example, navigating to work. Given the ubiquity of touch for BVI individuals, pandemic-related concerns around touch and the need to minimize contact with other people and public-facing surfaces impose significant challenges. Some of these COVID-19–related difficulties have been discussed previously,6 but the impacts of the coronavirus on spatial awareness and spatial behaviors by BVI people are still poorly understood. Given the importance of nonvisual spatial perception for this community, especially on the basis of touch, and that safe and efficient travel is critical for independence, any barrier represents a serious threat to the lives of millions of BVI people. Our focus here is on how COVID-19 limitations on touch and physical contact have led to unintended yet significant challenges to spatial perception, interpretation, and behavior for BVI individuals. These issues can be considered through the lens of spatial cognition, which is a broad field of interdisciplinary research that encompasses the knowledge and beliefs people have about the spatial properties of objects and events in the world, the manner that they explicitly acquire, mentally represent, and act upon this knowledge, and the spatial supports (eg, maps, simulations, language) used to represent spatial information.7, 8 A significant component of what supports spatiocognitive activities performed without vision is touch. Obvious applications of touch to environmental awareness and spatial cognition include detecting and identifying objects and localizing key features in the environment (eg, doorknobs, railings, and so on).9 However, touch is also an important input for directly supporting safe and efficient navigation. For instance, BVI people may use their long cane or foot to track the edge of a sidewalk, the feel of the tactile domes at an intersection to correctly orient when crossing the street, the feel of a distinct brick wall as a landmark, and myriad other tactile cues to maintain accurate orientation and safe navigation.10, 11 Success in these endeavors inevitably involves physical contact with many environmental elements and frequent proximity to other pedestrians.12 For BVI people, following appropriate social distancing behavior is particularly challenging because nonvisual sensing is not conducive to accurate detection or maintenance of a fixed 6-foot separation or an egocentric geo-fence. In most cases, BVI people are within this “bubble” when using their long cane and the ability to maintain any type of fixed boundary is inconsistent at best. Although the cane affords a traveler with the ability to swipe in a circle, thereby providing a “bubble of protection,” doing so is highly impractical. In normal use, the cane is used only as a forward-facing “probe,” with its field of view limited to an arc sweep of 90° to 100°, within a 3- to 5-foot range.13 Empirical data support that cane use is based on inconsistent sampling from this limited forward-facing region during travel.13 If dog guides are used, there is no training in place to maintain this separation. Indeed, service dogs are trained to utilize all possible space when guiding, meaning that they will often bring their handler within 1 or 2 feet of a passerby, especially when navigating in busy or crowded situations. The net result is that BVI travelers will generally, albeit unintentionally, end up much closer to surrounding pedestrians than their sighted peers, especially when navigating on busy streets, subway platforms, line queues, and so on. The following illustration may help elucidate some COVID-19–related navigation and spatial cognition challenges experienced by BVI people. Caitlin decides to walk to the nearby convenience store with her dog guide Sam. When she reaches the store, she identifies the handle by touch (with exposure regret) and goes inside. As she walks to the cooler in the back of the store, she inadvertently goes the “wrong” direction down the aisle as she cannot see (and is therefore completely unaware of) the newly demarcated directional arrows on the floor. Upon reaching the cooler, she feels for its handle (by necessity yet again) and then inside to find her beverage of choice, which she can recognize by the unique shape of the bottles. Approaching the check-out line, she is unable to see the floor markers indicating correct spatial separation and thus ends up much closer than she intends to a customer who is standing silently at the register who brusquely requests that she back off and stay behind the indicator, which she has no way of detecting. When making her purchase, she must touch the Plexiglas safety barrier that has been newly erected on the previously unobstructed counter to isolate the open area to place her 6-pack (an obvious massive exposure risk given the throughput of customers at this common point of purchase). Finally, as she reaches for her change, her hand makes inadvertent contact with the salesclerk's hand. Normally, most of these instances of physical contact, proximity, and movement behavior are everyday occurrences as a BVI person interacts with their world and are neither noteworthy nor problematic. However, in the coronavirus environment, many of these activities are potentially dangerous and put both Caitlin and those around her at greater risk. Her journey to the store is navigated primarily through touch and exploration with the hand. Although a sighted person may engage the door with an elbow (lowering risk), this is not practical if you cannot see where the handle is. Thus the hand is generally the most efficient and practical effector for supporting such behaviors, which makes sense as it is also one of the body regions with the highest tactile acuity.14, 15 However, it is precisely these types of public surfaces that are most likely to be vectors of COVID-19 and what people are advised to avoid contacting. That advice, although well intentioned and of little consequence to most sighted people, is not practical for BVI people. One solution is to wear gloves, which would allow for touching without direct skin contact. However, gloves represent a barrier between the skin and external stimulus, and such intervening materials can result in changes to tactile perception.16 Gloves may mask what is being felt and reduce tactile sensitivity, especially for discriminating high-resolution stimuli (ie, trying to read braille with gloves). Even for more general tasks, such as Caitlin experienced, most BVI people find gloves to be very distracting and disruptive to tactile perception, somewhat analogous to a sighted person walking around the world while wearing a pair of blur glasses or operating in a dense fog. Normally, instances of accidental contact, such as touching a salesclerk's hand or bumping into somebody in a line, are not problematic; a simple “excuse me” suffices. However, in the COVID-19 world, this type of accidental physical contact is frequently met with concern, fear, and sometimes hostility. Given that increased reliance on touch and inadvertent physical contact with others is the reality for BVI people, and that gloves are not a practical solution for this community to mitigate coronavirus health risks, one alternative is to use a large amount of hand sanitizer or to frequently wash the hands to maintain hygiene.17 However, this solution also presents challenges, as these methods when repeatedly deployed in short time windows lead to chapped finger pads that also limit sensitivity. It is analogous to the classic finger pruning that occurs when we spent too long in the bath as a child or take a few too many laps in the swimming pool. Indeed, even repeated, short duration exposure to water can negatively impact skin sensitivity.18 As BVI people will continue to use touch to experience their world, and the need to reduce physical contact will continue as an important form of coronavirus risk mitigation, a viable solution must be able to support both of these needs. There is a range of such solutions that could assist BVI people in the present situation. One obvious solution is to avoid exposure and increase isolation compensated by more services. Although this is potentially viable, it is not practical. Services are finite and unfortunately compromised in the present pandemic state, and like their sighted peers, BVI individuals are anxious to once again be able to go outside, walk around their neighborhood, and exercise agency over their life. A second solution is for BVI people to rely on a close contact with whom they already have exposure. This person could serve as a sighted guide when required, assist with getting groceries, or provide sighted assistance when needed during these unprecedented times. Although this approach could work in theory, it also is impractical as it means a BVI person must wait until their friend or family member is available, which is antithetical to independence because the process fosters reliance on others. The third and most promising solution employs the use of technological tools, called assistive technologies (ATs). Throughout the last century, the long cane and dog guide continue to be the most commonly used tools for mobility.19, 20 However, limitations of these traditional ATs, such as their small range of operation and limited field of view, has led to the development of many electronic ATs and travel aids (ETAs).21 Although these devices are often separated into those that help avoid hazards or that assist in orientation, the key features of a comprehensive AT solution are: (1) detecting obstacles in the travel path, (2) identifying travel surface information including texture and elevation discontinuities, (3) detecting objects bordering the travel path, (4) identifying distant objects and cardinal directions, (5) identifying landmark locations, and (6) providing sufficient spatial information to enable familiarity and mental mapping.22 In our current era of self-driving vehicles and automated route mapping, there is clear need to move beyond the long cane or dog guide used in conjunction with a potpourri of limited-use gadgets, aids, and apps. What is required is a paradigm shift focused on developing comprehensive tools that make safe mobility a reality for BVI travelers and that augment the existing primary mobility solutions with useful AT that provides complementary information supporting robust spatial awareness and cognition. One such platform is the VIS4ION system (Visually Impaired Smart Service System for Spatial Intelligence and Onboard Navigation). This platform (Figure 1) provides real-time situational and obstacle awareness in one's immediate environment, allowing individuals with visual impairment to travel more safely in three-dimensional (3D) space. VIS4ION remedies some of the cane's shortcomings, and further augments the ability of BVI persons to both maintain balance and to localize objects in their environment.23, 24 The system also provides robust networked features, which expands computational power through connectivity.9, 25-28 More specifically, VIS4ION is a mobile platform capable of real-time scene understanding with human-in-the-loop navigation assistance; the smart service system has four components: (1) a wearable backpack with several distinct distance and ranging/image sensors, which extract pertinent information about obstacles and the environment; (2) an embedded system with both computing and communication capability (inside backpack); (3) a haptic interface (waist strap) that communicates the spatial information computed from the sensory data to the end-user in real-time via an intuitive, ergonomic, and personalized vibrotactile belt (waist straps of book bag) positioned on the torso29-31; and (4) a head set that contains both binaural bone conduction speakers and a microphone for oral communication.23, 24 Users of the VIS4ION platform are alerted to environmental features of interest through the two human-machine-interface outputs: (1) audible messages delivered through bone conduction while leaving normal air-based audition intact; and (2) vibrotactile feedback whereby the scene that has been mapped is broken into a grid of segments and displayed to the end user in a crude pixelated form factor through the waist strap/belt. More specifically, the scene is decomposed into capture fields that correspond to the haptic interface in a spatiotopically preserved, intuitive, body-centered (eg, ocentric) fashion (Figure 2). Although this AT tool was clearly not developed for pandemic-related risk mitigation, the potential to double down on embedded technologies to combat COVID-19 is certainly present. In fact, the team is presently exploring computer vision-based approaches for solving a number of critical issues raised in the literature32, 33; although still under development, the wearable system is able to extract pertinent visual information from the user's surroundings, process the data through a series of parallel deep-learning techniques, and translate the results into pandemic-pertinent alerts and notifications. Spatial hazards are analyzed on three parameters: density (the population of the crowd), distance (from the user), and motion (relative to the user); subsequently risk is stratified for each parameter and an overall risk determination is rendered. Three maps are presented in Figure 3 to highlight these parameters: a crowd-density map (dens-map), distance map (dis-map), and motion map (motion-map). Higher density values are dark green in the density and lower density values are as In the distance greater distance values are dark and smaller distance values are as In the motion the a closer the BVI the the the a walks from the the the As in Figure an end user is a small of pedestrians in a and as the physical distance on what be for an appropriate social risk is and a risk is delivered by the spatial are based on the and environment. Although technological are not presently to with risk in the era of much of what has been developed could be with simple to the BVI In fact, many computer neural are robust for accurately detecting pedestrians and could provide and maintain a distance from others. to a of is and within given the sensing of many current embedded A more practical tool may be the we already have in our There are a of applications that provide sighted assistance through include and Although it has limited a guide may very well several of the in this present The is that limitations must be and clearly including but not being limited to compromised fields of view, often focused on service and development that often to the and of nonvisual interface In the coronavirus disease 19 (COVID-19) pandemic has created cataclysmic repercussions in while challenges and for those with particularly those with and visual impairment. distancing and the of touch are for but for those with visual they are For a viable solution to be it is critical that current ATs are and viable are put to the safety and agency of the BVI The simple is that the underlying problem is a it is a of information As a we must to more comprehensive solutions that to environmental and user that world is of also of the of