City-scale Datasets Research Articles

Distributed rate control of smart solar arrays with batteries

Continued advances in technology have led to falling costs and a dramatic increase in the aggregate amount of solar capacity installed across the world. A drawback of increased solar penetration is the potential for supply-demand mismatches in the grid due to the intermittent nature of solar generation. While energy storage can be used to mask such problems, we argue that there is also a need to explicitly control the rate of solar generation of each solar array in order to achieve high penetration while also handling supply-demand mismatches. To address this issue, we present the notion of smart solar arrays that can actively modulate their solar output based on the notion of proportional fairness. We present a decentralized algorithm based on Lagrangian optimization that enables each smart solar array to make local decisions on its fair share of solar power it can inject into the grid and then present a sense-broadcast-respond protocol to implement our decentralized algorithm into smart solar arrays. We also study the benefits of using energy storage when we rate control solar. To do so, we present a decentralized algorithm to charge and discharge batteries for each smart solar. Our evaluation on a city-scale dataset shows that our approach enables 2.6× more solar penetration while causing smart arrays to reduce their output by as little as 12.4%. By employing an adaptive gradient approach, our decentralized algorithm has 3 to 30× faster convergence. Finally, we demonstrate energy storage can help netmeter more solar energy while ensuring fairness and grid constraints are met.

New indicator for a comprehensive evaluation of building energy performance through spatial and temporal dimensions

Building functioning requires energy consumption. Traditional building energy performance (BEP) certification typically takes energy consumption as a single indicator, neglecting to evaluate the effect of energy application. Based on the definition of BEP and integrated consideration of energy consumption and service management, this study proposes a methodological path to construct a comprehensive performance indicator for buildings by evaluating energy consumption efficiency and amplitude of fluctuation using the XGBoost algorithm and objective assignment method. The indicator compares the deviation between actual and benchmark energy consumption in the spatial dimension and the amplitude of energy efficiency ratio fluctuations in the temporal dimension. Consequently, a comprehensive BEP evaluation model is thus coupled to form the comprehensive energy performance indicator (CEPI). The proposed approach is verified based on a city-scale dataset of over 2000 buildings, with CEPIi = 0.5288ECIi + 0.4712*MPIi for Chongqing. Compared with other evaluation systems, it is more intuitive, simple and robust, and can objectively reflect the rationality of energy use and provide a basis for scientific judgment.

VSLN: View‐aware sphere learning network for cross‐view vehicle re‐identification

Cross-view vehicle Reidentification (ReID) has attracted widespread attention as an increasingly important vision task in intelligent transportation and urban surveillance. Benefiting from Convolutional Neural Network (CNN), recent studies have promoted the development of vehicle ReID by extracting discriminative local features. However, two fundamental challenges of small interclass discrepancy caused by different views and large intraclass distance caused by similar appearance still hinder the performance of cross-view vehicle ReID. In this paper, a novel View-aware Sphere Learning Network (VSLN) is proposed to alleviate the above issues while maintaining the merits of CNN-based approaches to generate view-aware sphere-based features. First, a Sphere Feature Embedding Network (SFEN) is proposed to constrain the images into hypersphere for extracting sphere features. On the other hand, this study presents a sphere similarity triple loss to help SFEN concentrate more on robust and discriminative vehicle parts. Second, since the vehicle images are usually captured from different viewpoints, this study further extends SFEN by introducing a Vehicle Viewpoint Predictor (VVP) combined with global attention mechanism to enlarge the discrepancy of interclass and shorten the distance of intraclass. Moreover, a city-scale data set, named Vehicle from Different Viewpoints, containing image-level viewpoint labels, is collected for training VVP. As a result, the proposed VLSN can achieve 96.31% Top-1 accuracy and 79.46% Top-1 accuracy on VeRi-776 and VRIC data sets, respectively. Overall, extensive experimental results on two benchmark data sets show that the proposed VSLN outperforms state-of-the-art methods.

In-Route Task Selection in Spatial Crowdsourcing

Consider a city’s road network and a worker who is traveling on a given path from a starting point s to a destination d (e.g., from school or work to home) in said network. Consider further that there is a set of tasks in the network available to be performed, where each such task takes a certain amount of time to be completed and yields a positive reward if completed, and, finally, that the worker is willing to deviate from his/her path as long as the travel time to the selected tasks plus the time taken for completing them does not exceed a given time budget. We call this problem the In-Route Task Selection (IRTS) problem and consider two variants thereof. In the first one, named IRTS-SP, we assume that the worker only specifies s and d and he/she wants to consider alternative paths that deviate (cost-wise) as little as possible from the cost of the shortest path connecting s and d . In the second variant, named IRTS-PP, we assume that the worker has a preferred path from s to d and wants to travel along that one path for as long as possible. The latter is practically relevant in cases where the worker has a path other than the shortest one that is more desirable for non-objective reasons, e.g., availability of public transit, bicycle-friendliness or perceived safety. Common to both variants though, we assume that the worker wants to maximize the rewards collected by completing tasks. Clearly, there are now two conflicting criteria for the worker to contemplate when considering which tasks to perform: minimizing path deviation and maximizing collected reward. In this context, we investigate both IRTS variants using the skyline paradigm in order to obtain the set of non-dominated solutions w.r.t. the tradeoffs between earned rewards and deviation from either the cost of the shortest path, in the case of IRTS-SP, or the actual preferred path, in the case of IRTS-PP. Returning the skyline set of solutions to workers is of practical interest as it empowers them, e.g., it allows them to decide, at query time, which tasks suit them better. We propose exact and heuristic approaches in order to solve both variants of the IRTS problem. Our experiments, using real city-scale datasets, show that while the exact approaches serve as benchmarks, they do not scale due to the NP-hardness of the problems. The overall best heuristic approach, on the other hand, can solve relatively large instances of the IRTS problems within practical query processing time, e.g., at par with less effective greedy heuristics, while still producing very good approximate skyline sets, e.g., often yielding less than 10% relative error w.r.t. the exact solution.

Development of city buildings dataset for urban building energy modeling

Urban building energy modeling (UBEM) is becoming a proven tool to support energy efficiency programs for buildings in cities. Development of a city-scale dataset of the existing building stock is a critical step of UBEM to automatically generate energy models of urban buildings and simulate their performance. This study introduces data needs, data standards, and data sources to develop city building datasets for UBEM. First, a literature review of data needs for UBEM was conducted. Then, the capabilities of the current data standards for city building datasets were reviewed. Moreover, the existing public data sources from several pioneer cites were studied to evaluate whether they are adequate to support UBEM. The results show that most cities have adequate public data to support UBEM; however, the data are represented in different formats without standardization, and there is a lack of common keys to make the data mapping easier. Finally, a case study is presented to integrate the diverse data sources from multiple city departments of San Francisco. The data mapping process is introduced and discussed. It is recommended to use the unique building identifiers as the common keys in the data sources to simplify the data mapping process. The integration methods and workflow are applied to other U.S. cities for developing the city-scale datasets of their existing building stock, including San Jose, Los Angeles, and Boston.

Smartphone Sensing Meets Transport Data: A Collaborative Framework for Transportation Service Analytics

We advocate for and introduce TRANSense , a framework for urban transportation service analytics that combines participatory smartphone sensing data with city-scale transportation-related transactional data (taxis, trains, etc.). Our work is driven by the observed limitations of using each data type in isolation: (a) commonly-used anonymous city-scale datasets (such as taxi bookings and GPS trajectories) provide insights into the aggregate behavior of transport infrastructure, but fail to reveal individual-specific transport experiences (e.g., wait times in taxi queues); while (b) mobile sensing data can capture individual-specific commuting-related activities, but suffers from accuracy and energy overhead challenges due to usage artefacts and lack of appropriate sensing triggers. TRANSense demonstrates how a judicious fusion of such disparate data sources can overcome these challenges and offer novel insights. We detail two examples: (a) Taxi Service Analyzer that provides accurate detection of commuter queuing for taxis and estimates their wait time, by using taxi trip records to identify potential taxi locations with high demand and subsequently selectively triggering mobile sensing-based queuing analytics on nearby commuters; and (b) Subway Boarding Analyzer that identifies instances when passengers fail to board arriving trains, by first estimating train arrivals from temporal patterns of passenger egress at station gantries, and then using mobile sensing-based analysis of commuter movement behavior on platforms. Experiments with real-world datasets (from over 20,000 taxis and 1.7 million commuters in Singapore) show the power of this approach: the taxi service analyzer detects commuter queuing with over 90 percent accuracy with negligible energy overhead and estimates wait times with error margins below 15 percent, whereas the subway boarding analyzer can detect failed boarding events with a precision of over 90 percent (more than thrice what is achievable through purely mobile sensing).