We present methods and applications for the development of digital twins (DT) for urban traffic management. While the majority of studies on the DT focus on its “eyes,” which is the emerging sensing and perception like object detection and tracking, what really distinguishes the DT from a traditional simulator lies in its “brain,” the prediction and decision making capabilities of extracting patterns and making informed decisions from what has been seen and perceived. In order to add value to urban transportation management, DTs need to be powered by artificial intelligence and complement with low-latency high-bandwidth sensing and networking technologies, in other words, cyberphysical systems. This paper can be a pointer to help researchers and practitioners identify challenges and opportunities for the development of DTs; a bridge to initiate conversations across disciplines; and a road map to exploiting potentials of DTs for diverse urban transportation applications.

We propose a framework for adaptive data collection aimed at robust learning in multi-distribution scenarios under a fixed data collection budget. In each round, the algorithm selects a distribution source to sample from for data collection and updates the model parameters accordingly. The objective is to find the model parameters that minimize the expected loss across all the data sources. Our approach integrates upper-confidence-bound (UCB) sampling with online gradient descent (OGD) to dynamically collect and annotate data from multiple sources. By bridging online optimization and multi-armed bandits, we provide theoretical guarantees for our UCB-OGD approach, demonstrating that it achieves a minimax regret of $O(T^{\frac{1}{2}}(K\ln T)^{\frac{1}{2}})$ over $K$ data sources after $T$ rounds. We further provide a lower bound showing that the result is optimal up to a $\ln T$ factor. Extensive evaluations on standard datasets and a real-world testbed for object detection in smart-city intersections validate the consistent performance improvements of our method compared to baselines such as random sampling and various active learning methods.

This paper introduces PAVE (Pedestrian Awareness Via Edge analytics), a scalable real-time video analytics system that uses street cameras to enhance pedestrian safety while preserving their privacy. PAVE processes live camera streams on an edge server to track pedestrians and vehicles in real-time, predict vehicles' trajectories, and identify danger zones where pedestrians are present. The coordinates of these zones are sent to pedestrians' mobile devices via a custom iOS app, which locally determines if they are at risk without sharing any data with the edge server, hence preserving privacy. Moreover, anonymized metadata, including real-time location and speed/direction of pedestrians and vehicles, are visualized on a public map. PAVE's effectiveness was validated through deployment on the NSF COSMOS testbed, processing live video from cameras in diverse urban environments. Live field tests show that PAVE can alert at-risk pedestrians ~0.9 s before a vehicle reaches them. Through extensive profiling, we show that optimizing memory/compute configuration per pipeline stage can reduce latency by up to 10× compared to the default operating system configurations.

We showcase the workflow of PAVE (Pedestrian Awareness Via Edge analytics), a scalable system for real-time video analytics that leverages street cameras to improve pedestrians' safety while maintaining their privacy. PAVE distributes computation across edge servers and end-user mobile devices. Cameras' live streams are processed at the edge to forecast vehicles' trajectories and detect danger zones. Pedestrians' mobile devices then locally determine if the user is inside a danger zone and trigger timely alerts via a custom iOS app. In addition, anonymized metadata, such as pedestrian and vehicle positions, speeds, and directions, are aggregated and displayed on a public map for broader situational awareness. We evaluated PAVE's performance through implementation on the NSF COSMOS testbed's edge server while processing real-time video stream from cameras in diverse urban environments. Live field tests at an intersection in New York City show that PAVE can alert at-risk pedestrians about 0.9 s before a vehicle reaches them. With low-latency cameras, this lead time extends to around 1.6 s which is within the 1–2 s window pedestrians typically need to react.

Ensuring the safety of Vulnerable Road Users (VRUs) at intersections is crucial to enhancing urban traffic systems. This paper introduces a novel intelligent warning system specifically designed to increase the safety of VRUs crossing intersections. The proposed system leverages the COSMOS testbed to obtain real time vehicle information and employs Message Queuing Telemetry Transport (MQTT) as a standards-based messaging protocol for device communication and data transmission and utilizes a transformer model and Time To Collision (TTC) method to predict the collision. To validate the effectiveness and reliability of our intelligent alert system, we conducted comprehensive tests using the CARLA simulator, incorporating hardware in the loop simulation approach. The results demonstrate the potential for increased situational awareness and reduced risk factors associated with VRUs at intersections. Our work supports the integration of this intelligent alert system as a viable solution for reducing accidents and enhancing the overall safety of urban intersections in real time.

Recent advances in Visual Language Models (VLMs) have significantly enhanced video analytics. VLMs capture complex visual and textual connections. While Convolutional Neural Networks (CNNs) excel in spatial pattern recognition, VLMs provide a global context, making them ideal for tasks like complex incidents and anomaly detection. However, VLMs are much more computationally intensive, posing challenges for large-scale and real-time applications. This paper introduces EdgeCloudAI, a scalable system integrating VLMs and CNNs through edge-cloud computing. Edge-CloudAI performs initial video processing (e.g., CNN) on edge devices and offloads deeper analysis (e.g., VLM) to the cloud, optimizing resource use and reducing latency. We have deployed EdgeCloudAI on the NSF COSMOS testbed in NYC. In this demo, we will demonstrate EdgeCloudAI's performance in detecting user-defined incidents in real-time.

Blind and low-vision (BLV) people rely on GPS-based systems for outdoor navigation. GPS’s inaccuracy, however, causes them to veer off track, run into obstacles, and struggle to reach precise destinations. While prior work has made precise navigation possible indoors via hardware installations, enabling this outdoors remains a challenge. Interestingly, many outdoor environments are already instrumented with hardware such as street cameras. In this work, we explore the idea of repurposing existing street cameras for outdoor navigation. Our community-driven approach considers both technical and sociotechnical concerns through engagements with various stakeholders: BLV users, residents, business owners, and Community Board leadership. The resulting system, StreetNav, processes a camera’s video feed using computer vision and gives BLV pedestrians real-time navigation assistance. Our evaluations show that StreetNav guides users more precisely than GPS, but its technical performance is sensitive to environmental occlusions and distance from the camera. We discuss future implications for deploying such systems at scale.

Social distancing is an effective public health tool to reduce the spread of respiratory pandemics such as COVID-19. To analyze compliance with social distancing policies, we design two video-based pipelines for social distancing analysis, namely, automated video-based social distancing analyzer (Auto-SDA) and bird’s eye view social distancing analyzer (B-SDA). Auto-SDA is designed to measure social distancing using street-level cameras. To avoid privacy concerns of using street-level cameras, we further develop B-SDA, which uses bird’s eye view cameras, thereby preserving pedestrian’s privacy. We used the COSMOS testbed deployed in West Harlem, New York City (NYC), to evaluate both pipelines. In particular, Auto-SDA and B-SDA are applied on videos recorded by two of COSMOS cameras deployed on the 2nd floor (street-level) and 12th floor (bird’s eye view) of Columbia University’s Mudd building, looking at 120th St. and Amsterdam Ave. intersection, NYC. Videos are recorded before and during the peak of the pandemic, as well as after the vaccines became broadly available. The results represent the impact of social distancing policies on pedestrians’ social behavior. For example, the analysis shows that after the lockdown, less than 55% of the pedestrians failed to adhere to the social distancing policies, whereas this percentage increased to 65% after the vaccines’ availability. Moreover, after the lockdown, 0%–20% of the pedestrians were affiliated with a social group, compared to 10%–45% once the vaccines became available. The results also show that the percentage of face-to-face failures has decreased from 42.3% (prepandemic) to 20.7% (after the lockdown).

BACKGROUND: In the United States, over 12 000 home healthcare agencies annually serve 6+ million patients, mostly aged 65+ years with chronic conditions. One in three of these patients end up visiting emergency department (ED) or being hospitalized. Existing risk identification models based on electronic health record (EHR) data have suboptimal performance in detecting these high-risk patients. OBJECTIVES: To measure the added value of integrating audio-recorded home healthcare patient-nurse verbal communication into a risk identification model built on home healthcare EHR data and clinical notes. METHODS: This pilot study was conducted at one of the largest not-for-profit home healthcare agencies in the United States. We audio-recorded 126 patient-nurse encounters for 47 patients, out of which 8 patients experienced ED visits and hospitalization. The risk model was developed and tested iteratively using: (1) structured data from the Outcome and Assessment Information Set, (2) clinical notes, and (3) verbal communication features. We used various natural language processing methods to model the communication between patients and nurses. RESULTS: Using a Support Vector Machine classifier, trained on the most informative features from OASIS, clinical notes, and verbal communication, we achieved an AUC-ROC = 99.68 and an F1-score = 94.12. By integrating verbal communication into the risk models, the F-1 score improved by 26%. The analysis revealed patients at high risk tended to interact more with risk-associated cues, exhibit more "sadness" and "anxiety," and have extended periods of silence during conversation. CONCLUSION: This innovative study underscores the immense value of incorporating patient-nurse verbal communication in enhancing risk prediction models for hospitalizations and ED visits, suggesting the need for an evolved clinical workflow that integrates routine patient-nurse verbal communication recording into the medical record.

Blind and low-vision (BLV) people use GPS-based systems for outdoor navigation assistance, which provide instructions to get from one place to another. However, such systems do not provide users with real-time, precise information about their location and surroundings which is crucial for safe navigation. In this work, we investigate whether street cameras can be used to address aspects of navigation that BLV people still find challenging with existing GPS-based assistive technologies. We conducted formative interviews with six BLV participants to identify specific challenges they face in outdoor navigation. We discovered three main challenges: anticipating environment layouts, avoiding obstacles while following directions, and crossing noisy street intersections. To address these challenges, we are currently developing a street camera-based navigation system that provides real-time auditory feedback to help BLV users avoid obstacles, know exactly when to cross the street, and understand the overall layout of the environment. We close by discussing our evaluation plan.

OBJECTIVES: Patient-clinician communication provides valuable explicit and implicit information that may indicate adverse medical conditions and outcomes. However, practical and analytical approaches for audio-recording and analyzing this data stream remain underexplored. This study aimed to 1) analyze patients' and nurses' speech in audio-recorded verbal communication, and 2) develop machine learning (ML) classifiers to effectively differentiate between patient and nurse language. MATERIALS AND METHODS: Pilot studies were conducted at VNS Health, the largest not-for-profit home healthcare agency in the United States, to optimize audio-recording patient-nurse interactions. We recorded and transcribed 46 interactions, resulting in 3494 "utterances" that were annotated to identify the speaker. We employed natural language processing techniques to generate linguistic features and built various ML classifiers to distinguish between patient and nurse language at both individual and encounter levels. RESULTS: A support vector machine classifier trained on selected linguistic features from term frequency-inverse document frequency, Linguistic Inquiry and Word Count, Word2Vec, and Medical Concepts in the Unified Medical Language System achieved the highest performance with an AUC-ROC = 99.01 ± 1.97 and an F1-score = 96.82 ± 4.1. The analysis revealed patients' tendency to use informal language and keywords related to "religion," "home," and "money," while nurses utilized more complex sentences focusing on health-related matters and medical issues and were more likely to ask questions. CONCLUSION: The methods and analytical approach we developed to differentiate patient and nurse language is an important precursor for downstream tasks that aim to analyze patient speech to identify patients at risk of disease and negative health outcomes.

Crowded metropolises present unique challenges to the potential deployment of autonomous vehicles. Safety of pedestrians cannot be compromised and personal privacy must be preserved. Smart city intersections will be at the core of Artificial Intelligence (AI)-powered citizen-friendly traffic management systems for such metropolises. Hence, the main objective of this work is to develop an experimentation framework for designing applications in support of secure and efficient traffic intersections in urban areas. We integrated a camera and a programmable edge computing node, deployed within the COSMOS testbed in New York City, with an Eclipse sensiNact data platform provided by Kentyou. We use this pipeline to collect and analyze video streams in real-time to support smart city applications. In this demo, we present a video analytics pipeline that analyzes the video stream from a COSMOS' street-level camera to extract traffic/crowd-related information and sends it to a dedicated dashboard for real-time visualization and further assessment. This is done without sending the raw video, in order to avoid violating pedestrians' privacy.

Social distancing can reduce infection rates in respiratory pandemics such as COVID-19, especially in dense urban areas. Hence, we used the PAWR COSMOS wireless edge-cloud testbed in New York City to design and evaluate two different approaches for social distancing analysis. The first, \textbf{Auto}mated video-based \textbf{S}ocial \textbf{D}istancing \textbf{A}nalyzer (\textbf{Auto-SDA}), was designed to measure pedestrians compliance with social distancing protocols using street-level cameras. However, since using street-level cameras can raise privacy concerns, we also developed the \textbf{B}ird's eye view \textbf{S}ocial \textbf{D}istancing \textbf{A}nalyzer (\textbf{B-SDA}) which uses bird's eye view cameras, thereby preserving pedestrians' privacy. Both Auto-SDA and B-SDA consist of multiple modules. This demonstration illustrates the roles of these modules and their overall performance in evaluating the compliance of pedestrians with social distancing protocols. Moreover, we demonstrate applying Auto-SDA and B-SDA on videos recorded from cameras deployed on the 2nd and 12th floor of Columbia's Mudd building, respectively.

This paper focuses on COSMOS - Cloud enhanced Open Software defined MObile wireless testbed for city-Scale deployment. The COSMOS testbed is being deployed in West Harlem (New York City) as part of the NSF Platforms for Advanced Wireless Research (PAWR) program. It will enable researchers to explore the technology "sweet spot" of ultra-high bandwidth and ultra-low latency in the most demanding real-world environment. We describe the testbed's architecture, the design and deployment challenges, and the experience gained during the design and pilot deployment. Specifically, we describe COSMOS' computing and network architectures, the critical building blocks, and its programmability at different layers. The building blocks include software-defined radios, 28 GHz millimeter-wave phased array modules, optical transport network, core and edge cloud, and control and management software. We describe COSMOS' deployment phases in a dense urban environment, the research areas that could be studied in the testbed, and specific example experiments. Finally, we discuss our experience with using COSMOS as an educational tool.

The presence of multipaths leads to a loss of orthogonality between signals transmitted simultaneously on a wideband code-division multiple access (WCDMA) downlink. We derive general analytical expressions for the orthogonality factor (OF), which quantifies this loss of orthogonality, as a function of the instantaneous multipath fade realization of the channel. We show that the OF exhibits a significant temporal variation for the three channel profiles suggested in the WCDMA standard, namely, typical urban and rural areas and hilly terrain, which span a wide range of realistic cases. Moreover, its temporal variations and statistics vary significantly from one channel profile to another. Also, while the pulse shape and the number of RAKE fingers has only a marginal impact on the OF's statistics, the granularity in setting the finger positions has a considerable impact. The results of the work can be directly used for evaluation of the system performance of WCDMA cellular systems.

We explore the feasibility of designing an outdoor cellular network based on the IEEE 802.11 standard, which was developed originally for wireless local-area networks. For channels typical in cellular networks, we study the radio link power budget and, via simulation, the bit-error performance of three kinds of receiver: (1) the constrained RAKE, which is limited to a 1 /spl mu/s multipath span; (2) the full RAKE, which uses the full multipath channel information; (3) the ideal equalizer, the performance of which is represented by the matched filter bound. Our link budget reveals that the maximum cell radius in an outdoor 802.11 network ranges from 0.7 to 3 km, about half that supported by WCDMA and EDGE networks. For an RMS delay spread of 1 /spl mu/s, typical for urban-area cells of this size, our simulation results show that the conventional constrained RAKE receiver may yield a satisfactory performance. The improved receivers, however, yield 1-5 dB gain over the constrained implementation. Combining these results with those in a companion paper on the MAC protocol (see Leung, K.K. et al., ibid., p.595-599), we conclude that an 802.11 based cellular network with a cell radius of a few km is feasible.

We study the data performance of WCDMA systems using the downlink shared channel (DSCH) by investigating the impact of loading, rate adaptation and power control for the typical urban (TU) channel. We conclude that, using DSCH, high speed data transmissions can be achieved if the offered traffic load is well controlled. We discuss a heuristic rate adaptation algorithm that adapts the transmission rate based upon the perceived user performance in previous frames. We show that the performance of the rate adaptation is significantly better than that for the fixed-rate allocation policy.

The presence of multipath leads to a loss of orthogonality between the signals transmitted simultaneously on a WCDMA downlink. We derive general analytical expressions for the orthogonality factor (OF), which quantifies this loss of orthogonality. We show that the orthogonality factor exhibits a significant temporal variation for the three channel profiles suggested in the WCDMA standard, namely, typical urban, rural area, and hilly terrain, which span a wide range of realistic cases. Moreover, its temporal variations and statistics vary significantly from one channel profile to another. Also, while the pulse shape and the number of RAKE fingers has only a marginal impact on the OF's statistics, the granularity in setting the finger positions has a considerable impact. The results of the work can be directly used for evaluation of the system performance of WCDMA cellular systems.

We evaluate the performance of a previously proposed dynamic frequency hopping (DFH) when applied to cellular systems with a limited total bandwidth. We also illustrate a practical implementation for DFH deployment using network-assisted resource allocation (NARA). The performance evaluation is accomplished by system-level simulations of a system with 12 carriers and 1/1 frequency reuse, based on the EDGE-Compact specification. Voice-only circuit-switched operation is assumed. The fading channel, multicell interference, voice activity, and antenna sectorization are modeled. We present the performance of dynamic frequency hopping compared to random frequency hopping and fixed channel assignment by showing the distributions of the word error rates. The sensitivity to occupancy, Rayleigh fading assumptions, number of carriers, voice activity, and measurement errors are studied. We also compare the uplink and downlink performance. The results indicate that DFH can significantly improve the performance compared to random frequency hopping. For example, at a 2 % frame error rate with 90% coverage, the capacity improvement of DFH is almost 100% when compared with fixed channel assignment, and about 50% when compared to random frequency hopping. The amount of improvement for the uplink direction is smaller than the improvement for the downlink direction, especially for higher occupancies.

We investigate the performance of the WWW traffic due to the interactions between the transmission control protocols (TCP) and radio link protocols (RLPs) of various reliability. The study is done from the perspective of a cellular service operator, interested in maximizing the overall system performance. We implemented simulations of a spectrally-efficient cellular system based on the EDGE standard with numerous simultaneous users, each generating WWW packet traffic and operating the RENO TCP protocol. Due to the interference-limited nature of the system, the users experience over-the-air interactions similar to those of the congestion channel for which the TCP is suitable. The simulations indicate that tuning the retransmission persistence of the RLP protocol has a notable impact on the system performance. It is noted that the system performance when the TCP is used is not significantly worse when compared to the pure link layer performance.

We examine techniques for increasing spectral efficiency of cellular systems by using slow frequency hopping (FH) with dynamic frequency-hop (DFH) pattern adaptation. We first present analytical results illustrating the improvements in frequency outage probabilities obtained by DFH in comparison with random frequency hopping (RFH). Next, we show simulation results comparing the performance of various DFH and RFH techniques. System performance is expressed by cumulative distribution functions of codeword error rates. Systems that we study incorporate channel coding, interleaving, antenna diversity, and power control. Analysis and simulations consider the effects of path loss, shadowing, Rayleigh fading, cochannel interference, coherence bandwidth, voice activity, and occupancy. The results indicate that systems using DFH can support substantially more users than systems using RFH.

In this paper we evaluate the performance of dynamic frequency hopping (DFH) applied to cellular systems with a limited total bandwidth. The evaluation is done by system-level simulations of a system with 12 carriers and 1/1 frequency reuse, based on the EDGE-Compact system (EDGE-enhanced data rates in existing digital cellular systems). Voice-only circuit-switched operation is assumed. Fading channel, multi-cell interference, voice activity and antenna sectorization are modeled. We present the performance of dynamic frequency hopping compared to random frequency hopping and fixed channel assignment by showing distributions of word error rates. Sensitivity to occupancy, Rayleigh fading assumptions, number of carriers, voice activity and measurement errors are studied. The results indicate that DFH can significantly improve performance compared to random frequency hopping. For example, at a 2% frame error rate with 90% coverage, the capacity improvement of DFH is almost 100% when compared with fixed channel assignment, and about 50% when compared to random frequency hopping.

This paper reports on studies of the effects of the TCP/IP protocol header on the system performance of a spectrally efficient cellular system. Integrated cellular system and TCP-protocol simulations are used to examine the degradation in the effective data throughput and packet delay due to the header overhead. Performance improvements achievable by header compression are investigated. Simulations have been designed for an EDGE-like cellular system. System-wide performance is examined using cumulative distributions for data throughput, delay and packet success. For example, the data throughput of a micro-cellular system with Web traffic is improved about 10% when header compression from 40 to 2 bytes is implemented. More aggressive compression yielded minimal further improvements.

Spread spectrum (SS) multiple access techniques have been proposed for third generation broadband wireless access. We develop an analytical framework to quantify the effects of spreading bandwidth on SS systems operating in dense multipath environments in terms of the receiver performance, receiver complexity, and multipath channel parameters. In particular, we consider wide-sense stationary uncorrelated scattering (WSSUS) Gaussian channels with frequency-selective fading. The focus of the paper is to characterize the combined signal of the RAKE receiver fingers tracking the strongest multipath components. Closed form expressions for the mean and the variance of the total RAKE receiver output signal-to-noise ratio (SNR) are derived in terms of the number of RAKE fingers, spreading bandwidth, and multipath spread of the channel. The proposed problem is made analytically tractable by transforming the physical RAKE paths into the virtual path domain. A representative result indicates that for SS systems with 5 MHz signal bandwidth operating in a channel with constant power delay profile having 5 /spl mu/s spread, the average SNR gain from increasing the number of RAKE fingers from one to three is 3.8 dB and from three to five is 1.5 dB. Furthermore, the reduction in the variation of SNR is 1.1 dB and 0.4 dB for the same increments in the number of fingers.

New families of spread-spectrum codes are constructed, that are applicable to optical synchronous code-division multiple-access (CDMA) communications as well as to arbitrary-medium time-hopping synchronous CDMA communications. Proposed constructions are based on the mappings from integer sequences into binary sequences. The authors use the concept of number theoretic quadratic congruences and a subset of Reed-Solomon codes similar to the one utilized in the Welch-Costas frequency-hop (FH) patterns. The properties of the codes are as good as or better than the properties of existing codes for synchronous CDMA communications: both the number of code-sequences within a single code family and the number of code families with good properties are significantly increased when compared to the known code designs. Possible applications are presented. To evaluate the performance of the proposed codes, a new class of hit arrays called cyclical hit arrays is recalled, which give insight into the previously unknown properties of the few classes of number theoretic FH patterns. Cyclical hit arrays and the proposed mappings are used to determine the exact probability distribution functions of random variables that represent interference between users of a time-hopping or optical CDMA system. Expressions for the bit error probability in multi-user CDMA systems are derived as a function of the number of simultaneous CDMA system users, the length of signature sequences and the threshold of a matched filter detector. The performance results are compared with the results for some previously known codes.>

The problem of the algebraic construction of a particular family of optical codes for use in code-division multiple-access (CDMA) fiber-optic local area networks (LANs) is treated. The conditions that the code families have to satisfy when used in such systems are reviewed. The new codes are called quadratic congruence codes, and the construction of the corresponding sequences is based on the number-theoretic concept of quadratic congruences. It is shown that p-1 codes exist for every odd prime p and can serve as many as p-1 different users in the CDMA fiber-optic system. The codes belong to the family of optical orthogonal codes, their auto- and cross-correlation properties are established, and their performance is compared to that of the previous optical codes. Examples of the codes and examples of their auto- and cross-correlation functions are given.>

A method for the estimation of the parameters of multipath communication channels is developed. The method makes use of matched filtering, set-theoretic deconvolution and autoregressive modeling. Pulse compression waveforms are used as channel probing signals. The results of a feasibility study in which the proposed method is applied to specular multipath channels to estimate the multipath parameters at a number of different signal-to-noise ratios (SNR) are presented.>

A method for the estimation of the parameters of specular multipath communication channels is presented. The method takes advantage of the matched filtering, set-theoretic deconvolution and autoregressive modeling. Pulse compression waveforms such as chip and frequency-hop coding waveforms are used as channel problem signals. The proposed method, applied to a number of multipath channel models with closely spaced path components, produces high-resolution estimates of multipath parameters at signal-to-noise ratios higher than 5 dB.>