Denoising autoencoders for fast real-time traffic estimation on urban road networks

Ghosh, Soham; Asif, Muhammad; Wynter, Laura

doi:10.1109/cdc.2017.8264610

Cited by 8 publications

(4 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Autoencoder (AE) methods are a variety of artificial neural networks that learn an efficient encoding of unlabeled data in an unsupervised manner. In traffic prediction, AEs are primarily leveraged to impute or effectively compress data (Ghosh et al, 2017;Wang et al, 2018b), reducing its dimensionality while retaining the most vital information. The network design is relatively simple at the highest level, with two main portions: (1) an encoder that compresses the input (x) into fewer bits; and (2) a decoder that takes the sparse representation of the input (x ′ ) and outputs the original value.…”

Section: Basic Dnn Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Traffic prediction using artificial intelligence: Review of recent advances and emerging opportunities

Shaygan¹,

Meese²,

Li³

et al. 2022

Transportation Research Part C: Emerging Technologies

View full text Add to dashboard Cite

Section: Basic Dnn Modelsmentioning

confidence: 99%

“…imputation(Ghosh et al, 2017), data compression(Wang et al, 2018b), and feature extraction(Liu et al, 2019a). Specifically for traffic prediction, AEs are typically stacked in sequence to hierarchically extract the most important features from the traffic data.…”

mentioning

confidence: 99%

Traffic prediction using artificial intelligence: Review of recent advances and emerging opportunities

Shaygan¹,

Meese²,

Li³

et al. 2022

Transportation Research Part C: Emerging Technologies

View full text Add to dashboard Cite

“…As one type of machine learning, deep learning methods draw a lot of attention from academia and industry. Ghosh et al proposed a method consisting of the low-rank matrix and stacked denoising autoencoders (AEs), achieving promising imputation performance ( 35 ). However, it lacks the analysis of the performance on the different missing model.…”

Section: Related Workmentioning

confidence: 99%

Traffic Mosaic: Traffic Data Imputation Method Inspired by Image Super Resolution

Shuai

et al. 2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Missing data is a major problem in data-driven intelligent transportation systems. This study presents an imputation method, called TrafficMosaic, for effective imputation of lost traffic flow data. If we consider traffic data as a image, then the missing data in it can be regarded as a mosaic. The design of TrafficMosaic is inspired by super resolution techniques. First, a deep neural network (DNN) is employed to recover the lost data by a suite of techniques inspired from image super resolution. This DNN method combines the temporal relevance of historical flow with spatial correlation of current fragmented flow data, and uses the imputation result as input for recursive calculation. Second, a data-driven road network matrixing algorithm is proposed to mine location relations from trajectory data and reconstruct a road network flow into matrixes of road network flow pictures. Convolutional local calculation in a convolutional neural network is introduced to extract local spatial features of the road network, thereby reducing the computational complexity and improving the generalization ability. The final step is to extract the local temporal features of the flow data. Traffic flows at adjacent moments have strong temporal characteristics, and we splice several historical road network traffic pictures in chronological order so that our model can capture temporal information from the sequence. The experimental evaluation used a real automatic number plate recognition dataset for experiments, and the results show the effectiveness of TrafficMosaic for lost data imputation.

show abstract

“…When the communication capacity reduces, white Gaussian noise can increase the AEE greatly and results in low accuracy of traffic estimation. Therefore, it is necessary to add some denoising approach [35] to filter the noise from the collected data set, which is caused by uncorrelated factors, such as the parking vehicle along the roadside and other outlier measurements. However, it is observed that when the magnitude of noise is less than 20%, the proposed algorithm can still work well when the communication capacity is high.…”

Section: Performance Evaluationmentioning

confidence: 99%

Bandwidth-Aware Traffic Sensing in Vehicular Networks with Mobile Edge Computing

Kong

Dai

et al. 2019

Sensors

View full text Add to dashboard Cite

Traffic sensing is one of the promising applications to guarantee safe and efficient traffic systems in vehicular networks. However, due to the unique characteristics of vehicular networks, such as limited wireless bandwidth and dynamic mobility of vehicles, traffic sensing always faces high estimation error based on collected traffic data with missing elements and over-high communication cost between terminal users and central server. Hence, this paper investigates the traffic sensing system in vehicular networks with mobile edge computing (MEC), where each MEC server enables traffic data collection and recovery in its local server. On this basis, we formulate the bandwidth-constrained traffic sensing (BCTS) problem, aiming at minimizing the estimation error based on the collected traffic data. To tackle the BCTS problem, we first propose the bandwidth-aware data collection (BDC) algorithm to select the optimal uploaded traffic data by evaluating the priority of each road segment covered by the MEC server. Then, we propose the convex-based data recovery (CDR) algorithm to minimize estimation error by transforming the BCTS into an l 2 -norm minimization problem. Last but not the least, we implement the simulation model and conduct performance evaluation. The comprehensive simulation results verify the superiority of the proposed algorithm.

show abstract

Denoising autoencoders for fast real-time traffic estimation on urban road networks

Cited by 8 publications

References 26 publications

Traffic prediction using artificial intelligence: Review of recent advances and emerging opportunities

Traffic prediction using artificial intelligence: Review of recent advances and emerging opportunities

Traffic Mosaic: Traffic Data Imputation Method Inspired by Image Super Resolution

Bandwidth-Aware Traffic Sensing in Vehicular Networks with Mobile Edge Computing

Contact Info

Product

Resources

About