Your Selections

Intelligent and Connected Vehicles Symposium
Show Only Newly Added Content My Access Full Text Content


Complimentary Content
Journal Articles Magazine Articles Open Access TBMG Articles Tech Insights White Papers

File Formats

HTML (86) PDF (86)

Content Types

Technical Paper (86)



Automotive (82) Aerospace (3) Commercial Vehicle (2)


Mathematical models (31) Simulation and modeling (26) Autonomous vehicles (22) Vehicle drivers (22) Driver assistance systems (14) Optimization (13) Planning / sheduling (13) Radar (13) Research and development (12) Roads and highways (11) Sensors and actuators (11) Imaging and visualization (10) Neural networks (10) Yaw (10) CAD, CAM, and CAE (9) Cruise control (8) Automated Vehicles (7) Vehicle acceleration (7) Vehicle dynamics /flight dynamics (7) Control systems (6) Electric vehicles (6) Wheels (6) Driver behavior (4) Electric power (4) Frames (4) Hardware (4) Active safety systems (3) Artificial intelligence (AI) (3) Collision avoidance systems (3) Energy consumption (3) Hardware-in-the-loop (3) Logistics (3) Robotics (3) Simulators (3) Slip (3) Stability control (3) Test procedures (3) Traffic management (3) Vehicle handling (3) Architecture (2) Brake torque (2) Braking systems (2) Cartography (2) Collision warning systems (2) Comfort (2) Computer software and hardware (2) Connectors and terminals (2) Cutting (2) Data acquisition and handling (2) Education and training (2)


Huang, Libo (10) Bi, Xin (8) Bai, Jie (7) Deng, Weiwen (7) Chen, Hui (4) Chen, Sihan (4) Chen, Tao (4) Li, Fang (4) Wang, Lifang (4) Ando, Kazuya (3) Chen, Jiachen (3) Liu, Zhiguang (3) Nishimura, Yosuke (3) Wang, Shanshan (3) Wang, Yang Yang (3) Wu, Yan (3) Bian, Ning (2) Cao, Peng (2) Chen, Guangda (2) Feng, Rong (2) Gao, Feng (2) He, Rui (2) He, Xiangkun (2) He, Xiaolin (2) Hu, Chaowei (2) Ji, Xuewu (2) Joshi, Adit (2) Lan, Xiaoming (2) Lei, Ao (2) Li, Kai (2) Li, Wei (2) Li, Xin (2) Liu, Yulong (2) Ma, Zhixiong (2) Pan, Ding (2) Pei, Xiaofei (2) Tan, Bin (2) Wang, Jinsong (2) Wang, Yunpeng (2) Wang, Zhangyu (2) Wu, Jian (2) Xia, Qin (2) Xu, Zhijun (2) Yang, Bo (2) Yang, Cai (2) Yang, Kaiming (2) Yang, Shun (2) Yin, Yang (2) Yu, Guizhen (2) Zhu, Bing (2)


SAE (86)


Tongji University (14) Tongji Univ. (13) Jilin University (7) Beihang University (3) CATARC (3) Jilin Univ. (3) JTEKT Corporation (3) Tongji Univ (3) Wuhan University of Technology (3) Automotive Sensors Group (2) Chongqing University (2) Dongfeng Auto (2) Dongfeng Motor Corporation (2) Ford Motor Co., Ltd. (2) General Motors (2) Institute of Electrical Engineering,CAS (2) Key Laboratory of PEED, IEECAS (2) Key Laboratory of PEED,CAS (2) PilotD Automotive (Shanghai) Co. Ltd. (2) Shanghai International Automobile City (2) Tsinghua University (2) ABV- IIITM, Gwalior (1) Aerospace Hi-tech Holding Group CO.,LTD (1) ASCL, Jilin University (1) Beijing Union University (1) Beijing University of Posts & Telecom (1) Beijing Wanji Tech. Co., Ltd (1) Birmingham City Univ. (1) CAERI (1) CATARC Co., Ltd. (1) Chang'an University (1) Changan Automobile Company (1) Changan Automobile New Energy Research I (1) China Auto. Eng. Rsch Inst Co. Ltd. (1) China Automotive Eng Res Inst Co Ltd (1) China Automotive Engineering Research In (1) China Automotive Engrg Rsch Inst Co Ltd (1) China Automotive Technology & Research (1) Chongqin Changan Automobile Co., Ltd (1) Chongqing Changan Auto. Cor. Ltd. (1) Chongqing Changan Automobile Co. Ltd (1) Chongqing Jiaotong University (1) Chongqing Univ. (1) Chongqing University of Technology (1) Dongfeng Commercial Vehicle (1) Guangzhou Automobile Group (1) Harbin Institute of Technology (1) Hebei Univ. of Technology (1) Hebei University of Technology (1)


Intelligent and Connected Vehicles Symposium (86)

A Real-Time Traffic Light Detection Algorithm Based on Adaptive Edge Information

  • Beihang University - Guizhen Yu, Ao Lei, Honggang Li, Yunpeng Wang, Zhangyu Wang, Chaowei Hu
Published 2018-08-07 by SAE International in United States
Traffic light detection has great significant for unmanned vehicle and driver assistance system. Meanwhile many detection algorithms have been proposed in recent years. However, traffic light detection still cannot achieve a desirable result under complicated illumination, bad weather condition and complex road environment. Besides, it is difficult to detect multi-scale traffic lights by embedded devices simultaneously, especially the tiny ones. To solve these problems, this paper presents a robust vision-based method to detect traffic light, the method contains main two stages: the region proposal stage and the traffic light recognition stage. On region proposal stage, we utilize lane detection to remove partial background from the original image. Then, we apply adaptive canny edge detection to highlight region proposal in Cr color channel, where red or green color proposals can be separated easily. Finally, extract the enlarged traffic light RoI (Region of Interest) to classify. On traffic light recognition stage, a tinny but effective convolution neural network (CNN), named TLRNet, classifies each traffic light RoI into its own class. In fact, deep learning (DL) is bad for detecting small object in many fields, so we use region proposal stage to get RoI and classification by CNN to achieve a good result. We validate our method both on Laboratory for Intelligent and Safe Automobiles (LISA) Traffic Lights Dataset and video sequences captured from Beijing’s streets. The experimental results prove that the proposed method can achieve a good result for the multi-scales traffic lights in the TX1 embedded platform, and reach a real-time performance at 28fps.

Camera-Radar Data Fusion for Target Detection via Kalman Filter and Bayesian Estimation

  • Tongji Univ. - Zhexiang Yu, Jie Bai, Sihan Chen, Libo Huang, Xin Bi
Published 2018-08-07 by SAE International in United States
Target detection is essential to the advanced driving assistance system (ADAS) and automatic driving. And the data fusion of millimeter wave radar and camera could provide more accurate and complete information of targets and enhance the environmental perception performance. In this paper, a method of vehicle and pedestrian detection based on the data fusion of millimeter wave radar and camera is proposed to improve the target distance estimation accuracy. The first step is the targets data acquisition. A deep learning model called Single Shot MultiBox Detector (SSD) is utilized for targets detection in consecutive video frames captured by camera and further optimized for high real-time performance and accuracy. Secondly, the coordinate system of camera and radar are unified by coordinate transformation matrix. Then, the parallel Kalman filter is used to track the targets detected by radar and camera respectively. Since targets data provided by the camera and radar are different, different Kalman filters are designed to achieve the tracking process. Finally, the targets data are fused based on Bayesian Estimation. At first, several simulation experiments were designed to test and optimize the proposed method, then the real data was used to prove further. Through experiments, it shows that the measurement noise can be considerably reduced by Kalman filter and the fusion algorithm could improve the estimation accuracy.

Embedding CNN-Based Fast Obstacles Detection for Autonomous Vehicles

  • Beihang University - Chaowei Hu, Yunpeng Wang, Guizhen Yu, Zhangyu Wang, Ao Lei, Zhehua Hu
Published 2018-08-07 by SAE International in United States
Forward obstacles detection is one of the key tasks in the perception system of autonomous vehicles. The perception solution differs from the sensors and the detection algorithm, and the vision-based approaches are always popular. In this paper, an embedding fast obstacles detection algorithm is proposed to efficiently detect forward diverse obstacles from the image stream captured by the monocular camera. Specifically, our algorithm contains three components. The first component is an object detection method using convolution neural networks (CNN) for single image. We design a detection network based on shallow residual network, and an adaptive object aspect ratio setting method for training dataset is proposed to improve the accuracy of detection. The second component is a multiple object tracking method based on correlation filter for the adjacent images. Based on precise detection result, we use multiple correlation filters to track multiple objects in every adjacent frame, and a multi-scale tracking region algorithm is applied to improve the tracking accuracy at the same time. The third component is fusing the detection method and tracking method based on parallel processing, which can significantly increase the average processing rate for the image stream or video in embedded platform. Besides, our algorithm is tested on KITTI dataset as well as our own dataset, and the experimental results illustrate that our algorithm has high precision and robustness. Meanwhile, we test our algorithm on a popular embedded platform - NVIDIA Jetson TX1, and the average processing rate is approximately 17 fps, which satisfies the high processing speed requirements of autonomous vehicles.

Driver Risk Perception Model under Critical Cut-In Scenarios

  • Tongji University - Xuehan Ma, Zhiwei Feng, Xichan Zhu, Zhixiong Ma
Published 2018-08-07 by SAE International in United States
In China Cut-in scenarios are quite common on both highway and urban road with heavy traffic. They have a potential risk of rear-end collision. When facing a cutting in vehicle, driver tends to brake in most case to reduce collision risk. The timing and dynamic characteristics of brake maneuver are indicators of driver subjective risk perception. Time to collision (TTC) and Time Headway (THW) demonstrate objective risk. This paper aims at building a model quantitatively revealing the relationship between drivers’ subjective risk perception and objective risk. A total of 66 valid critical Cut-in cases was extracted from China-FOT, which has a travel distance of about 130 thousand miles. It is found that under Cut-in scenarios, driver tended to brake when the cutting in vehicle right crossing line. This time point was defined as initial brake time. Brake strength and brake speed were taken to describe brake maneuver. Average brake pressure (ABP) and acceleration at initial brake time indicated brake strength. Brake pressure change rate (BPCR) and longitudinal jerk (derivative of acceleration) at initial brake time indicated brake speed. Analytic Hierarchy Process and Fuzzy Comprehensive Evaluation Method were adopted to obtain an integrated subjective risk perception indicator D. Critical cases were divided into 3 groups by distance of within 5 m, from 5 to 15 m and over 15 m. Within the distance of 5 m, D was linear with 1/THW. Within the distance of from 5 to 15 m, D was linear with 1/TTC. Within the distance of over 15 m, both 1/THW and 1/TTC have linear relationship with D.

A Localization System for Autonomous Driving: Global and Local Location Matching Based on Mono-SLAM

  • Tongji Univ. - Zhijun Xu, Sihan Chen, Jie Bai, Libo Huang, Xin Bi
Published 2018-08-07 by SAE International in United States
The utilization of the SLAM (Simultaneous Localization and Mapping) technique was extended from the robotics to the autonomous vehicles for achieving the positioning. However, SLAM cannot obtain the global position of the vehicle but a relative one to the start. For sake of this, a fast and accurate system was proposed to obtain both the local position and the global position of vehicles based on mono-SLAM which realized the SLAM by using monocular camera with a lower cost and power consumption. Firstly, the rough latitude and longitude of current position was obtained by using common GPS without differential signal. Then, the Mono-SLAM operated on the consecutive video frames to generate the localization and local trajectory and its accuracy was further improved by utilizing the IMU information. After that, a piece of Map centered in the rough position obtained by common GPS was downloaded from the Open Street Map. Finally, a searching process in the downloaded Map was executed by using chamfer matching algorithm to find a piece of path matched with the constructed trajectory. Consequently, the global position of the vehicle was obtained and the accumulated error can be decreased with cyclical searching. In the test, the performance of this proposed system outperforms current approaches in global location and its error was less than 5 meters, indicating in parallel the potential that mono-SLAM can bring to the global localization task.

Critical Driving Scenarios Extraction Optimization Method Based on China-FOT Naturalistic Driving Study Database

  • Tongji University - Yufan Zeng, Xichan Zhu, Zhixiong Ma, Xiaoyu Sun
Published 2018-08-07 by SAE International in United States
Due to the differences in traffic situations and traffic safety laws, standards for extraction of critical driving scenarios (CDSs) vary from different countries and areas around the world. To maintain the characteristic variables under the Chinese typical CDSs, this paper uses the three-layer detection method to extract and detect CDSs in the Natural Driving Data from China-FOT project which executing under the real traffic situation in China. The first layer of detection is mainly based on the feature distributions which deviate from normal driving situations. These distributions associated with speed and longitudinal acceleration/lateral acceleration/yaw rate also quantify the critical levels classification. The second layer of detection based on the rate of brake pressure (Pressure peak/Time difference) and the relevant variables to TTC’s (Time to Collision) trigger, Pressure peak means the maximum value on brake pressure curve, Time difference means the difference between Pressure peak time and Hard breaking time (Time when driver starts to make emergency brake). The second layer could make corrections to the critical levels. The third layer of detection considers the effect of vehicle speed and make quantification of critical levels. The results show the accuracy (ACC) of detection under three-layer method makes greater optimization compared to other methods which analyze single variable. After the first two layer detections ACC achieves 69.71% while after the third layer detection ACC achieves 85.10%, 780 CDSs are extracted from these data. The results of this paper could provide a basis for the classification of CDSs from Natural Driving Data in China and causation mechanism of CDSs.

Application Oriented Testcase Generation for Validation of Environment Perception Sensor in Automated Driving Systems

  • PilotD Automotive (Shanghai) Co. Ltd. - Peng Cao
  • Tongji University - Libo Huang
Published 2018-08-07 by SAE International in United States
Validation is one of the main challenges in development of automated driving systems (ADS). Due to the complexity of these systems and the various influence factors on their functional safety, current testcase generation methods can hardly guarantee the completeness and effectivity of the validation on system level. Separate validation of system components is a way to make system approval possible. In this paper, an approach is presented to generate deductively testcases for the validation of the environment perception sensors, which are the most essential components of ADS. This approach is originated from the model-based testing method, which is commonly used to validate software-based systems and extended by considering various external influence factors as follows: By modeling and analyzing applications in ADS, application oriented usecases of perception sensors are first derived. Based on a classification of perception sensor errors, the sensor error types that are critical for each usecase are identified. Meanwhile, based on sensor working principle, the correspondence between external influence factors and each sensor error type are summarized in a morphological box. By combining the factors, which can “stimulate” a certain sensor error type that is critical for usecases, testcases can be generated. As an example, adaptive cruise control (ACC) system is analyzed as application in level 3 system instead of level 1 function in reality. This paper presents a structured deduction of testcases to make a complete validation of perception sensor for ADS possible. Some possibilities of testcase reduction and an outlook for further development and utilization are presented as well.

Path Planning Algorithm of Intelligent Vehicle Based on Improved Visibility Graphs

  • Chongqing Jiaotong University - Hanbing Wei
Published 2018-08-07 by SAE International in United States
Presently, the visibility graphs algorithm is mainly applied for path planning of indoor mobile robot. It only considers the constraints such as travelling time and move distance. The road lane and vehicle dynamics constraints are not deal with usually. In this paper, a local path planning algorithm based on improved visibility graphs is proposed for intelligent vehicle on structured road. First, free state space (FSS) is established based on ago-vehicle state, road lane and traffic condition for permitting ago-vehicle move safely. In FSS, the vehicle’s maneuver in preview distance can be inferred and the local target point can be designated. Next, sampling points is created in FSS. Combined with local target point, initial point and sampling points, road network can be generated consequently. Then, the approachable path in the road network are evaluated by constrains of the Euclidean distance and the vehicle dynamics constraints. In this way the unique shortest path satisfying the constraints are generated as an optimal path. Finally, the efficiency and performance of the algorithm are validated by simulation results. The comparison results with RRT algorithm show that, the planned path achieved by improved visibility graphs has shorter travel distance and lower curvature, which shows more superior advantage for the path planning of intelligent vehicle than RRT.

Analysis of Human Machine Interaction Program in Lane Keeping Assist System Based on Field Test

  • Tongji University - Yilin Yan
Published 2018-08-07 by SAE International in United States
Lane-keeping assist system (LKA) alerts the driver or intervenes in the driving when the vehicle deviates from the lane. But its effect is highly dependent on the driver’s acceptance. Distance to Lane Crossing (DTLC) and Time to Lane Crossing (TTLC) are two important factors to consider the danger level of the scenario, which are also two references for drivers to make decisions. At present, most of the functional design standards are based on these values, while they often differ for different vehicle movements.

Research on the Radiated Immunity Test Methods of ADAS Functions in Intelligent Vehicles

  • Vehicle Engineering Institute of CQUT - Jin Jia
  • CAERI - Biao Li
  • Show More
Published 2018-08-07 by SAE International in United States
According to the ISO 11451-2 [1], the immunity performance of passenger cars and commercial vehicle shall be validated to electrical disturbances from off-vehicle radiation sources. Typical vehicle electrical or electronic systems such as double flash, monitor, or audio and entertainments systems have been validated through a standardized immunity test process. As the development of ADAS (Automated Driver Assistance System) functions in the vehicle technology, the reliability and accuracy of these functions in the complex electromagnetic fields must be validated. Radar and Camera or Lidar sensor based functions of AEB (Automated Emerge Brake), ACC (Automated Cruise Control) or LDW (Lane Departure Warning) might be invalid, especially when these sensors were illuminated by electrical disturbances in the real word. This work presents a novel test methodology of whole-vehicle immunity to off-board electrical disturbances, considering the main ADAS functions. Target simulator and virtue driving-environment are designed to trigger AEB, ACC and LDW functions in an anechoic chamber. Under standardized off-board radiated sources, the reliability of ADAS functions in a real vehicle are invested.