|
| Author and year | Methods | Objective | Limitations |
|
| Wang et al. [49] 2009 | Spatial analysis approach | To scout the strike of congestion in traffic on the prevalence of distinct accidents in road | To cruise the effectuate of congestions at junctions on smash-up |
| Rempe [36] 2016 | Clustering algorithm | To decree congestion clusters that furnishes an allusive quantum of flexibility to confront the covenants for distinct applications | The composed method must be enforced and tested in an online form of a forecast system of traffic |
| Song et al. [51] 2017 | Convolutional neural network | To prognosticate the speed in traffic and analogize the performances with the existing prognosticate models | The multiple submodels are persistent |
| Hao et.al [50] 2017 | (1) Sparse mobile crowd
(2) Integrating traffic state model, emission model, and dispersion model | To forge a system that furnishes suggestions to the respective officials to alleviate the exposure of air pollutant | A system must be demonstrated to visualize the dispersion and emission of traffic |
| Onyeneke et al. [35] 2018 | (1) Systematic random sampling
(2) Purposive sampling
(3) Simple linear regression
(4) Correlation | To audit the effects that is independent of withal fabricating or importing conveyance in the United States | Perception of performance metric is persistent |
| Hebert et al. [52] 2019 | (1) Balanced random forest algorithm
(2) XG boost algorithm | To nurture high-resolution accident prognosticate model using big data analytics | More features and a dataset with population density are persistent for delicate performance |
| Wei et al. [55] 2019 | Autoencoder long short-term memory (AE-LSTM) | To ameliorate the prognosticate accuracy in flow of traffic | Simple spatial patterns and time patterns are only premeditated in this study |
| Du et al. [53] 2020 | Hybrid multimodel deep learning framework for traffic flow forecasting (HMDLF | To portent the short-term traffic flow | Confound in competent collection of data on accidents and baroque weather events in shorter time period |
| Moses and Parvathi [54] 2020 | (1) Support vector regression
(2) Mean squared error
(3) Linear regression model
(4) Decision tree learning | To prefabricate an efficient model for prognosticating the traffic volume and for effectuating out the hidden insights in vehicular stirrings | Arduous in inferring the traffic problem in real time |
| Bang and Lee [25] 2020 | (1) Vector-based mobility prediction model
(2) TDMA-based VANET | To avert blending or access collision between conveyances by prognosticating the delinquent stirring position and direction of each conveyance | Radical access and merging fracases intervene due to conveyance patterns in the movement and conditions of traffic |
| Ranjan et al. [41] 2020 | (1) Convolutional neural network (CNN)
(2) Long short-term memory (LSTM)
(3) Transpose CNN | To prognosticate the congestion level by grasping the chronological ramification of input images | Abundance of resources and computational time are indulgent in ascertaining the background area |
| Sellami and Alaya [56] 2021 | Self-adaptive multikernel clustering for urban VANET (SAMNET) | To strike the unpredictable density and to stipulate a certain balance load | Adopting this approach is complex in road scenarios, and optimizing the performance is difficult |
|
