A Multi-task Learning Framework for Road Attribute Updating via Joint Analysis of Map Data and GPS Traces

Yin, Yifang; Varadarajan, Jagannadan; Wang, Guanfeng; Wang, Xueou; Sahrawat, Dhruva; Zimmermann, Roger; Ng, See-Kiong

doi:10.1145/3366423.3380021

Cited by 8 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There have been a plethora of researches on mobility analysis based on GPS trajectories, many of which focus on intra-city vehicle data analysis [1, 5, 10, 12, 20, 23, 28, 29, 36-38, 41, 42]. Among these, a good number of works have focused on identification and characterization of various point of interests over a route, which can impact the mobility of the vehicles, such as point of intersections [5], various road attributes [13,40,42,47], mobility patterns [24], traffic congestion [14,19,29], significant locations on a route [23], and so on. A handful of these works also consider intra-city and public bus travels [12,13,19,23,24,30], where the GPS data has been collected either through a vehicle-mounted sensor or from smartphone crowdsensing.…”

Section: Related Workmentioning

confidence: 99%

Exploiting Multi-modal Contextual Sensing for City-bus's Stay Location Characterization: Towards Sub-60 Seconds Accurate Arrival Time Prediction

Mandal,

Karmakar,

Chatterjee

et al. 2021

Preprint

View full text Add to dashboard Cite

Intelligent city transportation systems are one of the core infrastructures of a smart city. The true ingenuity of such an infrastructure lies in providing the commuters with real-time information about citywide transports like public buses, allowing her (him) to pre-plan the travel. However, providing prior information for transportation systems like public buses in real-time is inherently challenging because of the diverse nature of different stay-locations that a public bus stops. Although straightforward factors stay duration, extracted from unimodal sources like GPS, at these locations look erratic, a thorough analysis of public bus GPS trails for 720km of bus-travels at the city of Durgapur, a semi-urban city in India, reveals that several other fine-grained contextual features can characterize these locations accurately. Accordingly, we develop BuStop, a system for extracting and characterizing the stay locations from multi-modal sensing using commuters' smartphones. Using this multi-modal information BuStop extracts a set of granular contextual features that allow the system to differentiate among the different stay-location types. A thorough analysis of BuStop using the collected in-house dataset indicates that the system works with high accuracy in identifying different stay locations like regular bus-stops, random ad-hoc stops, stops due to traffic congestion stops at traffic signals, and stops at sharp turns. Additionally, we also develop a proof-of-concept setup on top of BuStop to analyze the potential of the framework in predicting expected arrival time, a critical piece of information required to pre-plan travel, at any given bus-stop. Subsequent analysis of the PoC framework, through simulation over the test dataset, shows that characterizing the stay-locations indeed helps make more accurate arrival time predictions with deviations less than 60 seconds from the ground-truth arrival time.CCS Concepts: • Human-centered computing → Smartphones; Ubiquitous computing.

show abstract

Section: Related Workmentioning

confidence: 99%

Exploiting Multi-modal Contextual Sensing for City-bus's Stay Location Characterization: Towards Sub-60 Seconds Accurate Arrival Time Prediction

Mandal,

Karmakar,

Chatterjee

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Computer scientists and statistical physicists often adopt urban transportation networks as tractable, real‐world systems that can be well represented by graphs. OSMnx has been used accordingly to generate input graphs and feature sets for methodological research in machine learning and network algorithms (Yin et al., 2020; Young & Eccles, 2020). Ren, Cheng, and Zhang (2019) model Chengdu’s street network then predict traffic flow with a deep spatiotemporal residual neural network.…”

Section: Empirical Street Network Science With Osmnxmentioning

confidence: 99%

The right tools for the job: The case for spatial science tool‐building

Boeing

2020

Transactions in GIS

View full text Add to dashboard Cite

Do I need to know the precise polygonal geometries of Los Angeles and the University of Southern California (USC) to assert that the latter is within the former? No. My mind contains no such precise geometric model of points and lines, yet I know that USC is in Los Angeles. When humans reason with the real world, they focus on its objects, relations, and processes-rather than starting with geometry-because these are the keys to

show abstract

“…The idea of collecting spatial data indirectly, finding proxies for their characteristics, and amplifying data from other domains to serve GIScience are not new, e.g. Yin et al (2020) analyse movement trajectories to infer attributes of roads, Chen et al (2021b) mine social media data to map and understand amenities, Lines & Basiri (2021) exploit obstructions in satellite signals to reconstruct the vertical extent of buildings, Wu & Biljecki (2022) map buildings from street networks, Milojevic-Dupont et al (2020) infer the heights of buildings by developing a regression model that predicts them from the characteristics of the footprint and surrounding context, and Delmelle & Nilsson (2021) assess the ability of using property listing text for neighbourhood type prediction. However, to the extent of our knowledge, the potential of real estate data in building and amenity data acquisition remains uninvestigated despite their abundance, which is the key contribution of this paper.…”

Section: Introductionmentioning

confidence: 99%

Mining real estate ads and property transactions for building and amenity data acquisition

Chen

Biljecki

2022

Urban Info

View full text Add to dashboard Cite

Acquiring spatial data of fine and dynamic urban features such as buildings remains challenging. This paper brings attention to real estate advertisements and property sales data as valuable and dynamic sources of geoinformation in the built environment, but unutilised in spatial data infrastructures. Given the wealth of information they hold and their user-generated nature, we put forward the idea of real estate data as an instance of implicit volunteered geographic information and bring attention to their spatial aspect, potentially alleviating the challenge of acquiring spatial data of fine and dynamic urban features. We develop a mechanism of facilitating continuous acquisition, maintenance, and quality assurance of building data and associated amenities from real estate data. The results of the experiments conducted in Singapore reveal that one month of property listings provides information on 7% of the national building stock and about half of the residential subset, e.g. age, type, and storeys, which are often not available in sources such as OpenStreetMap, potentially supporting applications such as 3D city modelling and energy simulations. The method may serve as a novel means to spatial data quality control as it detects missing amenities and maps future buildings, which are advertised and transacted before they are built, but it exhibits mixed results in identifying unmapped buildings as ads may contain errors that impede the idea.

show abstract

A Multi-task Learning Framework for Road Attribute Updating via Joint Analysis of Map Data and GPS Traces

Cited by 8 publications

References 19 publications

Exploiting Multi-modal Contextual Sensing for City-bus's Stay Location Characterization: Towards Sub-60 Seconds Accurate Arrival Time Prediction

Exploiting Multi-modal Contextual Sensing for City-bus's Stay Location Characterization: Towards Sub-60 Seconds Accurate Arrival Time Prediction

The right tools for the job: The case for spatial science tool‐building

Mining real estate ads and property transactions for building and amenity data acquisition

Contact Info

Product

Resources

About