Deep learning models for traffic flow prediction in autonomous vehicles: A review, solutions, and challenges

Miglani, Arzoo; Kumar, Neeraj

doi:10.1016/j.vehcom.2019.100184

Cited by 177 publications

(85 citation statements)

References 104 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For training, they used different inductive learning models such as Decision Trees, k-Nearest Neighbor, Naive Bayes, and Bayes Net and evaluated their approach using five-fold cross-validation. Authors in [40][41][42][43][44][45][46] discussed the network protocol to use and process the sensors and wearable devices data over the network. The research [13] proposed an assisted approach that helps the participants to live healthily.…”

Section: Literature Reviewmentioning

confidence: 99%

Analyzing the Effectiveness and Contribution of Each Axis of Tri-Axial Accelerometer Sensor for Accurate Activity Recognition

Javed

Sarwar

Khan

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

Recognizing human physical activities from streaming smartphone sensor readings is essential for the successful realization of a smart environment. Physical activity recognition is one of the active research topics to provide users the adaptive services using smart devices. Existing physical activity recognition methods lack in providing fast and accurate recognition of activities. This paper proposes an approach to recognize physical activities using only2-axes of the smartphone accelerometer sensor. It also investigates the effectiveness and contribution of each axis of the accelerometer in the recognition of physical activities. To implement our approach, data of daily life activities are collected labeled using the accelerometer from 12 participants. Furthermore, three machine learning classifiers are implemented to train the model on the collected dataset and in predicting the activities. Our proposed approach provides more promising results compared to the existing techniques and presents a strong rationale behind the effectiveness and contribution of each axis of an accelerometer for activity recognition. To ensure the reliability of the model, we evaluate the proposed approach and observations on standard publicly available dataset WISDM also and provide a comparative analysis with state-of-the-art studies. The proposed approach achieved 93% weighted accuracy with Multilayer Perceptron (MLP) classifier, which is almost 13% higher than the existing methods.2 of 18 emotional issues, and depression as well. Physical fitness can be tracked and analyzed by monitoring daily life physical activities.Physical activity recognition was initiated back in 2004 using on-body sensors. Researchers in [2] used the accelerometer's annotated data to recognize the physical activities. They made an Android-based system that collects raw data from the accelerometer and applied machine learning algorithms to predict physical activities. Authors in [3] recognized six basic activities, i.e., walking, jogging, sitting, standing, upstairs and downstairs. Authors in [4] used on-body sensors for activity recognition but found that it is very difficult to carry the sensors all the time. Many authors suggested that a smartphone is a non-obtrusive option for activity recognition [3,[5][6][7][8][9].The smartphone is playing a vital role in modern life. It provides services and applications such as health monitoring, early-stage disease detection, sports analysis, fitness tracking, and behavior analysis. Android-based smartphones have a built-in motion sensor that provides accurate and precise acceleration readings against physical activities. In early conditions, dedicated sensors were used for activity recognition. There exist several techniques for physical activity recognition such as on-body obtrusive and non-obtrusive sensors [10][11][12][13]. Non-obtrusive sensors are used in smart homes and smartphones. In smart homes, different motion and door sensors are installed at different locations and the primary objective is to recognize...

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Analyzing the Effectiveness and Contribution of Each Axis of Tri-Axial Accelerometer Sensor for Accurate Activity Recognition

Javed

Sarwar

Khan

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Deep reinforcement learning provides the ability to output continuous action using DDPG and SAC for driving behavior [ 3 ]. Both DDPG and SAC can be further enhanced when instead of being trained on raw input, they are trained on outputs obtained through VAE with pre-defined loss functions [ 29 ]. For instance, the steering actions in a vehicle fluctuate a lot when the agent is trying to maintain its position in a lane or while making a turn [ 30 ].…”

Section: Literature Reviewmentioning

confidence: 99%

Policy-Gradient and Actor-Critic Based State Representation Learning for Safe Driving of Autonomous Vehicles

Gupta

Khwaja

Anpalagan

et al. 2020

Sensors

View full text Add to dashboard Cite

In this paper, we propose an environment perception framework for autonomous driving using state representation learning (SRL). Unlike existing Q-learning based methods for efficient environment perception and object detection, our proposed method takes the learning loss into account under deterministic as well as stochastic policy gradient. Through a combination of variational autoencoder (VAE), deep deterministic policy gradient (DDPG), and soft actor-critic (SAC), we focus on uninterrupted and reasonably safe autonomous driving without steering off the track for a considerable driving distance. Our proposed technique exhibits learning in autonomous vehicles under complex interactions with the environment, without being explicitly trained on driving datasets. To ensure the effectiveness of the scheme over a sustained period of time, we employ a reward-penalty based system where a negative reward is associated with an unfavourable action and a positive reward is awarded for favourable actions. The results obtained through simulations on DonKey simulator show the effectiveness of our proposed method by examining the variations in policy loss, value loss, reward function, and cumulative reward for `VAE+DDPG’ and `VAE+SAC’ over the learning process.

show abstract

“…Omer, Rofè, and Lerman (2015) assert additional features that affect pedestrian traffic flow, namely tourist sites and public transportation stations. Although various techniques are used to predict traffic flow, mainly implementing machine‐learning algorithms (Miglani & Kumar, 2019; Mihaita, Li, He, & Rizoiu, 2019), these mostly refer to vehicular traffic flow. To the best of our knowledge, none have tackled this problem using OSM data.…”

Section: Literature Overviewmentioning

confidence: 99%

Machine‐learning prediction models for pedestrian traffic flow levels: Towards optimizing walking routes for blind pedestrians

Cohen

Dalyot

2020

Transactions in GIS

View full text Add to dashboard Cite

show abstract

Deep learning models for traffic flow prediction in autonomous vehicles: A review, solutions, and challenges

Cited by 177 publications

References 104 publications

Analyzing the Effectiveness and Contribution of Each Axis of Tri-Axial Accelerometer Sensor for Accurate Activity Recognition

Analyzing the Effectiveness and Contribution of Each Axis of Tri-Axial Accelerometer Sensor for Accurate Activity Recognition

Policy-Gradient and Actor-Critic Based State Representation Learning for Safe Driving of Autonomous Vehicles

Machine‐learning prediction models for pedestrian traffic flow levels: Towards optimizing walking routes for blind pedestrians

Contact Info

Product

Resources

About