Lateral Motion Prediction of On-Road Preceding Vehicles: A Data-Driven Approach

Wang, Chen; Delport, Jacques; Wang, Yan

doi:10.3390/s19092111

Cited by 9 publications

(7 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chen Wang et al [10] have recently recommended a two-stage data-driven method. The dataset was created with the help of a safety pilot model development programme, from which summary statistics and Fast Fourier transform algorithms were used to extract features and labels.…”

Section: A Classification Based Approachesmentioning

confidence: 99%

PredictV: A Vehicle Prediction Scheme to Circumvent Occluded Frames

et al. 2022

View full text Add to dashboard Cite

There are many methods to overcome traffic congestion on the roads, but occlusion is still there in most methods, so it is a dire need of the time that researchers have to look into this matter. In the rural and urban areas, heavy congestion on the roads has become the leading cause of occlusion. The PredictV method works on the prediction principle based on existing values and is proved one of the naval approaches for this problem. This scheme uses blob detection for the first frame detection and predicts other vehicles based on percentage increment in the different parameters to identify the vehicles. For improving the quality of the sample, data cleaning has been included in this work. With the help of this approach, congestion has been avoided, and this method prefers to use predicted points to detect vehicles on the frame. The suggested approach is implemented via a MATLAB simulator. It is tested on a large dataset containing 7152 frames of 6 different videos from the Urban Tracker and KoPer datasets. In total, there are 46876 vehicles present on the frame at first, and the existing methods have detected only 58% and 73% of vehicles, whereas the detection rate is 82% with this suggested approach. The occlusion rate is only 17% on average, which has been reduced via this proposed approach which was 24% and 43% earlier than the existing one. The performance of this suggested method has been tested based on performance and results, which are pretty impressive that is almost 99.74%. In short, the outcome of this prediction technique has been improved now, and the effectiveness has been observed too.

show abstract

Section: A Classification Based Approachesmentioning

confidence: 99%

PredictV: A Vehicle Prediction Scheme to Circumvent Occluded Frames

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The current study is a step forward to use a deep reinforcement learning (DRL) to model safe autonomous driving behavior. Our results have been compared to other data-driven models based on the accuracy, maximum error-free drive-time before following too closely, and maximum error-free drive-time while overtaking [25]. These state-of-the-art developments in deep learning and reinforcement learning were the principal motivation behind this work.…”

Section: Motivation and Objectivementioning

confidence: 99%

Safe Driving Of Autonomous Vehicles Through Improved Deep Reinforcement Learning

Gupta¹

2021

Preprint

View full text Add to dashboard Cite

In this thesis, we propose an environment perception framework for autonomous driving using deep reinforcement learning (DRL) that exhibits learning in autonomous vehicles under complex interactions with the environment, without being explicitly trained on driving datasets. Unlike existing techniques, our proposed technique takes the learning loss into account under deterministic as well as stochastic policy gradient. We apply DRL to object detection and safe navigation while enhancing a self-driving vehicle’s ability to discern meaningful information from surrounding data. For efficient environmental perception and object detection, various Q-learning based methods have been proposed in the literature. Unlike other works, this thesis proposes a collaborative deterministic as well as stochastic policy gradient based on DRL. Our technique is a combination of variational autoencoder (VAE), deep deterministic policy gradient (DDPG), and soft actor-critic (SAC) that adequately trains a self-driving vehicle. In this work, we focus on uninterrupted and reasonably safe autonomous driving without colliding with an obstacle or steering off the track. We propose a collaborative framework that utilizes best features of VAE, DDPG, and SAC and models autonomous driving as partly stochastic and partly deterministic policy gradient problem in continuous action space, and continuous state space. To ensure that the vehicle traverses the road over a considerable period of time, we employ a reward-penalty based system where a higher negative penalty is associated with an unfavourable action and a comparatively lower positive reward is awarded for favourable actions. We also examine the variations in policy loss, value loss, reward function, and cumulative reward for ‘VAE+DDPG’ and ‘VAE+SAC’ over the learning process.

show abstract

“…The effects of reset control on the alleviation of the rise time, settling time, and overshoot limitations are explored for a lane change maneuver under a set of demanding design conditions to guarantee a suitable ride quality and a swift response [24]. A two-stage data-driven approach is proposed to classify driving patterns of surrounding vehicles, using Gaussian mixture models (GMM) [25]. Vehicles’ short-term lateral motions are predicted based on real-world vehicle mobility data, where several critical kinetic features and higher-order kinematic variables are utilized.…”

Section: Related Workmentioning

confidence: 99%

Conditional Artificial Potential Field-Based Autonomous Vehicle Safety Control with Interference of Lane Changing in Mixed Traffic Scenario

Gao

Yan

Yang

et al. 2019

Sensors

View full text Add to dashboard Cite

Car-following is an essential trajectory control strategy for the autonomous vehicle, which not only improves traffic efficiency, but also reduces fuel consumption and emissions. However, the prediction of lane change intentions in adjacent lanes is problematic, and will significantly affect the car-following control of the autonomous vehicle, especially when the vehicle changing lanes is only a connected unintelligent vehicle without expensive and accurate sensors. Autonomous vehicles suffer from adjacent vehicles’ abrupt lane changes, which may reduce ride comfort and increase energy consumption, and even lead to a collision. A machine learning-based lane change intention prediction and real time autonomous vehicle controller is proposed to respond to this problem. First, an interval-based support vector machine is designed to predict the vehicles’ lane change intention utilizing limited low-level vehicle status through vehicle-to-vehicle communication. Then, a conditional artificial potential field method is used to design the car-following controller by incorporating the lane-change intentions of the vehicle. Experimental results reveal that the proposed method can estimate a vehicle’s lane change intention more accurately. The autonomous vehicle avoids collisions with a lane-changing connected unintelligent vehicle with reliable safety and favorable dynamic performance.

show abstract

Lateral Motion Prediction of On-Road Preceding Vehicles: A Data-Driven Approach

Cited by 9 publications

References 25 publications

PredictV: A Vehicle Prediction Scheme to Circumvent Occluded Frames

PredictV: A Vehicle Prediction Scheme to Circumvent Occluded Frames

Safe Driving Of Autonomous Vehicles Through Improved Deep Reinforcement Learning

Conditional Artificial Potential Field-Based Autonomous Vehicle Safety Control with Interference of Lane Changing in Mixed Traffic Scenario

Contact Info

Product

Resources

About