1 Authorities in Montgomery Township,Pennsylvania,have introduced wavy lane patterns on some streets in an effort to slow down traffic in the area.2 Driving along Grays Lane in Montgomery Township for the first time m...1 Authorities in Montgomery Township,Pennsylvania,have introduced wavy lane patterns on some streets in an effort to slow down traffic in the area.2 Driving along Grays Lane in Montgomery Township for the first time must be challenging.Thats because the regular lane patterns have been replaced by wavy and zig⁃zag lines that look like they were painted by a drunk.But they are wavy by design.According to Montgomery Township officials,the unusual patterns were thought as the best solution to discouraging speeding on some of the streets.Police sources told local media outlets that the“traffic⁃calming measures”were installed in response to numerous complaints about certain streets being used as“speedways”.展开更多
In the field of autonomous driving,the task of reliably and accurately detecting lane markings poses a significant and complex challenge.This study presents a lane recognition model that employs an encoder-decoder arc...In the field of autonomous driving,the task of reliably and accurately detecting lane markings poses a significant and complex challenge.This study presents a lane recognition model that employs an encoder-decoder architecture.In the encoder section,we develop a feature extraction framework that operates concurrently with attention mechanisms and convolutional layers.We propose a spatial axis attention framework that integrates spatial information transfer regulated by gating units.This architecture places a strong emphasis on long-range dependencies and the spatial distribution of images.Furthermore,we incorporate multi-scale convolutional layers to extract intricate features from the images.The two sets of feature maps are concatenated and subsequently transformed into an input sequence for the decoder,with the lane marking coordinates considered as a target sequence for coordinate generation.This decoder can directly segment multiple lane markings,eliminating the need for additional post-processing algorithms,thereby significantly streamlining the lane recognition process.The proposed method demonstrates a high degree of accuracy in recognizing lane markings and exhibits robust capabilities in differentiating between occlusions and objects resembling lanes.It shows exceptional performance on the TuSimple and CULane datasets.展开更多
This paper addresses the lane-keeping control problem for autonomous ground vehicles subject to input saturation and uncertain system parameters.An enhanced adaptive terminal sliding mode based prescribed performance ...This paper addresses the lane-keeping control problem for autonomous ground vehicles subject to input saturation and uncertain system parameters.An enhanced adaptive terminal sliding mode based prescribed performance control scheme is proposed,which enables the lateral position error of the vehicle to be kept within the prescribed performance boundaries all the time.This is achieved by firstly introducing an improved performance function into the controller design such that the stringent initial condition requirements can be relaxed,which further allows the global prescribed performance control result,and then,developing a multivariable adaptive terminal sliding mode based controller such that both input saturation and parameter uncertainties are handled effectively,which further ensures the robust lane-keeping control.Finally,the proposed control strategy is validated through numerical simulations,demonstrating its effectiveness.展开更多
Lane line detection is a fundamental step in applications like autonomous driving and intelligent traffic monitoring. Emerging applications today have higher requirements for accurate lane detection. In this paper, we...Lane line detection is a fundamental step in applications like autonomous driving and intelligent traffic monitoring. Emerging applications today have higher requirements for accurate lane detection. In this paper, we present a precise information extraction algorithm for lane lines. Specifically, with Gaussian Mixture Model(GMM), we solved the issue of lane line occlusion in multi-lane scenes. Then, Progressive Probabilistic Hough Transform(PPHT) was used for line segments detection. After K-Means clustering for line segments classification, we solved the problem of extracting precise information that includes left and right edges as well as endpoints of each lane line based on geometric characteristics. Finally, we fitted these solid and dashed lane lines respectively. Experimental results indicate that the proposed method performs better than the other methods in both single-lane and multi-lane scenarios.展开更多
A new vision-based long-distance lane perception and front vehicle location method was developed for decision making of full autonomous vehicles on highway roads,Firstly,a real-time long-distance lane detection approa...A new vision-based long-distance lane perception and front vehicle location method was developed for decision making of full autonomous vehicles on highway roads,Firstly,a real-time long-distance lane detection approach was presented based on a linear-cubic road model for two-lane highways.By using a novel robust lane marking feature which combines the constraints of intensity,edge and width,the lane markings in far regions were extracted accurately and efficiently.Next,the detected lane lines were selected and tracked by estimating the lateral offset and heading angle of ego vehicle with a Kalman filter,Finally,front vehicles were located on correct lanes using the tracked lane lines,Experiment results show that the proposed lane perception approach can achieve an average correct detection rate of 94.37% with an average false positive detection rate of 0.35%,The proposed approaches for long-distance lane perception and front vehicle location were validated in a 286 km full autonomous drive experiment under real traffic conditions.This successful experiment shows that the approaches are effective and robust enough for full autonomous vehicles on highway roads.展开更多
In this paper, the speed gradient （SG） model is extended to describe the traffic flow on two-lane freeways. Terms related to lane change are added into the continuity equations and velocity dynamic equations. The em...In this paper, the speed gradient （SG） model is extended to describe the traffic flow on two-lane freeways. Terms related to lane change are added into the continuity equations and velocity dynamic equations. The empirically observed two-lane phenomena, such as lane usage inversion and lane change rate versus density, are reproduced by extended SG model. The local cluster effect is also investigated by numerical simulations.展开更多
A multilane extension of the single-lane anisotropic continuum model （GK model） developed by Gupta and Katiyar for traffic flow is discussed with the consideration of the coupling effect between the vehicles of diff...A multilane extension of the single-lane anisotropic continuum model （GK model） developed by Gupta and Katiyar for traffic flow is discussed with the consideration of the coupling effect between the vehicles of different lanes in the instantaneous traffic situation and the lane-changing effect. The conditions for securing the linear stability of the new model are presented. The shock and the rarefaction waves, the local cluster effect and the phase transition are investigated through simulation experiments with the new model and are found to be consistent with the diverse nonlinear dynamical phenomena observed in a real traffic flow. The analysis also focuses on empirically observed two- lane phenomena, such as lane usage inversion and the density dependence of the number of lane changes. It is shown that single-lane dynamics can be extended to multilane cases without changing the basic properties of the single-lane model. The results show that the new multilane model is capable of explaining some particular traffic phenomena and is in accordance with real traffic flow.展开更多
To explore the potential capacity of dual-fight-turn lanes at signalized intersections under mixed traffic conditions, we defined two conflict zones between right turn vehicles and through bicycle corresponding to dif...To explore the potential capacity of dual-fight-turn lanes at signalized intersections under mixed traffic conditions, we defined two conflict zones between right turn vehicles and through bicycle corresponding to different right turn flows from dual-right-turn lanes. Relationships between the arrival rate of bicycle group at each conflict zone and the saturation flow rate of right turn movement were investigated. A model based on gap acceptance theory was adopted to estimate the capacity of dual-right-turn lanes under mixed traffic conditions. An analysis was carried out using the collected data from three four-leg signalized intersections in Beijing, China, where the dual-right-turn lanes were used. In addition, we also discussed the patterns of bicycle lane in the urban area of Beijing, and classified it based on its characteristics in use. It is concluded that the two lanes of dual-fight-turn lanes produce different capacities under mixed traffic conditions, and the analysis on scenarios of dual-right-turn movement traversing bicycle traffic plays a key role in explaining the different capacity performance of the two right turn lanes. Error analysis of the model indicated that the model was rational.展开更多
A visual object-oriented software for lane following on intelligent highway system (IHS) is proposed. According to object-oriented theory, 3 typical user services of self-check, transfer of human driving and automatic...A visual object-oriented software for lane following on intelligent highway system (IHS) is proposed. According to object-oriented theory, 3 typical user services of self-check, transfer of human driving and automatic running and abnormal information input from the sensors are chosen out. In addition, the functions of real-time display, information exchanging interface, determination and operation interweaving in the 3 user services are separated into 5 object-oriented classes. Moreover, the 5 classes are organized in the visual development environment. At last, experimental result proves the validity and reliability of the control application.展开更多
The essential requirement for precise localization of a self-driving car is a lane-level map which includes road markings(RMs).Obviously,we can build the lane-level map by running a mobile mapping system(MMS)which is ...The essential requirement for precise localization of a self-driving car is a lane-level map which includes road markings(RMs).Obviously,we can build the lane-level map by running a mobile mapping system(MMS)which is equipped with a high-end 3D LiDAR and a number of high-cost sensors.This approach,however,is highly expensive and ineffective since a single high-end MMS must visit every place for mapping.In this paper,a lane-level RM mapping system using a monocular camera is developed.The developed system can be considered as an alternative to expensive high-end MMS.The developed RM map includes the information of road lanes(RLs)and symbolic road markings(SRMs).First,to build a lane-level RM map,the RMs are segmented at pixel level through the deep learning network.The network is named RMNet.The segmented RMs are then gathered to build a lane-level RM map.Second,the lane-level map is improved through loop-closure detection and graph optimization.To train the RMNet and build a lane-level RM map,a new dataset named SeRM set is developed.The set is a large dataset for lane-level RM mapping and it includes a total of 25157 pixel-wise annotated images and 21000 position labeled images.Finally,the proposed lane-level map building method is applied to SeRM set and its validity is demonstrated through experimentation.展开更多
A robust lane detection and tracking system based on monocular vision is presented in this paper. First, the lane detection algorithm can transform raw images into top view images by inverse perspective mapping （ IPM...A robust lane detection and tracking system based on monocular vision is presented in this paper. First, the lane detection algorithm can transform raw images into top view images by inverse perspective mapping （ IPM）, and detect both inner sides of the lane accurately from the top view im- ages. Then the system will turn to lane tracking procedures to extract the lane according to the infor- mation of last frame. If it fails to track the lane, lane detection will be triggered again until the true lane is found. In this system, 0-oriented Hough transform is applied to extract candidate lane mark- ers, and a geometrical analysis of the lane candidates is proposed to remove the outliers. Additional- ly, vanishing point and region of interest（ROI） dynamically planning are used to enhance the accura- cy and efficiency. The system was tested under various road conditions, and the result turned out to be robust and reliable.展开更多
文摘1 Authorities in Montgomery Township,Pennsylvania,have introduced wavy lane patterns on some streets in an effort to slow down traffic in the area.2 Driving along Grays Lane in Montgomery Township for the first time must be challenging.Thats because the regular lane patterns have been replaced by wavy and zig⁃zag lines that look like they were painted by a drunk.But they are wavy by design.According to Montgomery Township officials,the unusual patterns were thought as the best solution to discouraging speeding on some of the streets.Police sources told local media outlets that the“traffic⁃calming measures”were installed in response to numerous complaints about certain streets being used as“speedways”.
基金support was provided by Ministry of Education Higher Education Industry University Research Innovation Fund Special Project.If there are other authors,they declare that they have no known competing financial interests or personal relationships that could have appeared to influence thework reported in this paper.
文摘In the field of autonomous driving,the task of reliably and accurately detecting lane markings poses a significant and complex challenge.This study presents a lane recognition model that employs an encoder-decoder architecture.In the encoder section,we develop a feature extraction framework that operates concurrently with attention mechanisms and convolutional layers.We propose a spatial axis attention framework that integrates spatial information transfer regulated by gating units.This architecture places a strong emphasis on long-range dependencies and the spatial distribution of images.Furthermore,we incorporate multi-scale convolutional layers to extract intricate features from the images.The two sets of feature maps are concatenated and subsequently transformed into an input sequence for the decoder,with the lane marking coordinates considered as a target sequence for coordinate generation.This decoder can directly segment multiple lane markings,eliminating the need for additional post-processing algorithms,thereby significantly streamlining the lane recognition process.The proposed method demonstrates a high degree of accuracy in recognizing lane markings and exhibits robust capabilities in differentiating between occlusions and objects resembling lanes.It shows exceptional performance on the TuSimple and CULane datasets.
基金supported in part by the National Key Research and Development Program of China under Grant 2023YFA1011803in part by Natural Science Foundation of Chongqing,China under Grant CSTB2023NSCQ-MSX0588+2 种基金in part by the Fundamental Research Funds for the Central Universities,China under Grant 2023CDJKYJH047in part by the National Natural Science Foundation of China under Grant 62273064,Grant 61991400,Grant 61991403,Grant 61933012,Grant 62250710167,Grant 62203078in part by Innovation Support Program for International Students Returning to China under Grant cx2022016.
文摘This paper addresses the lane-keeping control problem for autonomous ground vehicles subject to input saturation and uncertain system parameters.An enhanced adaptive terminal sliding mode based prescribed performance control scheme is proposed,which enables the lateral position error of the vehicle to be kept within the prescribed performance boundaries all the time.This is achieved by firstly introducing an improved performance function into the controller design such that the stringent initial condition requirements can be relaxed,which further allows the global prescribed performance control result,and then,developing a multivariable adaptive terminal sliding mode based controller such that both input saturation and parameter uncertainties are handled effectively,which further ensures the robust lane-keeping control.Finally,the proposed control strategy is validated through numerical simulations,demonstrating its effectiveness.
基金supported by the National Nature Science Foundation of China under Grant No.61502226the Jiangsu Provincial Transportation Science and Technology Project No.2017X04the Fundamental Research Funds for the Central Universities
文摘Lane line detection is a fundamental step in applications like autonomous driving and intelligent traffic monitoring. Emerging applications today have higher requirements for accurate lane detection. In this paper, we present a precise information extraction algorithm for lane lines. Specifically, with Gaussian Mixture Model(GMM), we solved the issue of lane line occlusion in multi-lane scenes. Then, Progressive Probabilistic Hough Transform(PPHT) was used for line segments detection. After K-Means clustering for line segments classification, we solved the problem of extracting precise information that includes left and right edges as well as endpoints of each lane line based on geometric characteristics. Finally, we fitted these solid and dashed lane lines respectively. Experimental results indicate that the proposed method performs better than the other methods in both single-lane and multi-lane scenarios.
基金Project(90820302) supported by the National Natural Science Foundation of China
文摘A new vision-based long-distance lane perception and front vehicle location method was developed for decision making of full autonomous vehicles on highway roads,Firstly,a real-time long-distance lane detection approach was presented based on a linear-cubic road model for two-lane highways.By using a novel robust lane marking feature which combines the constraints of intensity,edge and width,the lane markings in far regions were extracted accurately and efficiently.Next,the detected lane lines were selected and tracked by estimating the lateral offset and heading angle of ego vehicle with a Kalman filter,Finally,front vehicles were located on correct lanes using the tracked lane lines,Experiment results show that the proposed lane perception approach can achieve an average correct detection rate of 94.37% with an average false positive detection rate of 0.35%,The proposed approaches for long-distance lane perception and front vehicle location were validated in a 286 km full autonomous drive experiment under real traffic conditions.This successful experiment shows that the approaches are effective and robust enough for full autonomous vehicles on highway roads.
文摘In this paper, the speed gradient （SG） model is extended to describe the traffic flow on two-lane freeways. Terms related to lane change are added into the continuity equations and velocity dynamic equations. The empirically observed two-lane phenomena, such as lane usage inversion and lane change rate versus density, are reproduced by extended SG model. The local cluster effect is also investigated by numerical simulations.
文摘A multilane extension of the single-lane anisotropic continuum model （GK model） developed by Gupta and Katiyar for traffic flow is discussed with the consideration of the coupling effect between the vehicles of different lanes in the instantaneous traffic situation and the lane-changing effect. The conditions for securing the linear stability of the new model are presented. The shock and the rarefaction waves, the local cluster effect and the phase transition are investigated through simulation experiments with the new model and are found to be consistent with the diverse nonlinear dynamical phenomena observed in a real traffic flow. The analysis also focuses on empirically observed two- lane phenomena, such as lane usage inversion and the density dependence of the number of lane changes. It is shown that single-lane dynamics can be extended to multilane cases without changing the basic properties of the single-lane model. The results show that the new multilane model is capable of explaining some particular traffic phenomena and is in accordance with real traffic flow.
文摘To explore the potential capacity of dual-fight-turn lanes at signalized intersections under mixed traffic conditions, we defined two conflict zones between right turn vehicles and through bicycle corresponding to different right turn flows from dual-right-turn lanes. Relationships between the arrival rate of bicycle group at each conflict zone and the saturation flow rate of right turn movement were investigated. A model based on gap acceptance theory was adopted to estimate the capacity of dual-right-turn lanes under mixed traffic conditions. An analysis was carried out using the collected data from three four-leg signalized intersections in Beijing, China, where the dual-right-turn lanes were used. In addition, we also discussed the patterns of bicycle lane in the urban area of Beijing, and classified it based on its characteristics in use. It is concluded that the two lanes of dual-fight-turn lanes produce different capacities under mixed traffic conditions, and the analysis on scenarios of dual-right-turn movement traversing bicycle traffic plays a key role in explaining the different capacity performance of the two right turn lanes. Error analysis of the model indicated that the model was rational.
文摘A visual object-oriented software for lane following on intelligent highway system (IHS) is proposed. According to object-oriented theory, 3 typical user services of self-check, transfer of human driving and automatic running and abnormal information input from the sensors are chosen out. In addition, the functions of real-time display, information exchanging interface, determination and operation interweaving in the 3 user services are separated into 5 object-oriented classes. Moreover, the 5 classes are organized in the visual development environment. At last, experimental result proves the validity and reliability of the control application.
基金This work was supported by the Industry Core Technology Development Project,20005062Development of Artificial Intelligence Robot Autonomous Navigation Technology for Agile Movement in Crowded Space,funded by the Ministry of Trade,industry&Energy(MOTIE,Republic of Korea).
文摘The essential requirement for precise localization of a self-driving car is a lane-level map which includes road markings(RMs).Obviously,we can build the lane-level map by running a mobile mapping system(MMS)which is equipped with a high-end 3D LiDAR and a number of high-cost sensors.This approach,however,is highly expensive and ineffective since a single high-end MMS must visit every place for mapping.In this paper,a lane-level RM mapping system using a monocular camera is developed.The developed system can be considered as an alternative to expensive high-end MMS.The developed RM map includes the information of road lanes(RLs)and symbolic road markings(SRMs).First,to build a lane-level RM map,the RMs are segmented at pixel level through the deep learning network.The network is named RMNet.The segmented RMs are then gathered to build a lane-level RM map.Second,the lane-level map is improved through loop-closure detection and graph optimization.To train the RMNet and build a lane-level RM map,a new dataset named SeRM set is developed.The set is a large dataset for lane-level RM mapping and it includes a total of 25157 pixel-wise annotated images and 21000 position labeled images.Finally,the proposed lane-level map building method is applied to SeRM set and its validity is demonstrated through experimentation.
基金Supported by the National Natural Science Foundation of China(51005019)
文摘A robust lane detection and tracking system based on monocular vision is presented in this paper. First, the lane detection algorithm can transform raw images into top view images by inverse perspective mapping （ IPM）, and detect both inner sides of the lane accurately from the top view im- ages. Then the system will turn to lane tracking procedures to extract the lane according to the infor- mation of last frame. If it fails to track the lane, lane detection will be triggered again until the true lane is found. In this system, 0-oriented Hough transform is applied to extract candidate lane mark- ers, and a geometrical analysis of the lane candidates is proposed to remove the outliers. Additional- ly, vanishing point and region of interest（ROI） dynamically planning are used to enhance the accura- cy and efficiency. The system was tested under various road conditions, and the result turned out to be robust and reliable.
