Search Results

The driver monitoring system (DMS) plays an essential role in reducing traffic accidents caused by human errors due to driver distraction and fatigue. The vision-based DMS has been the most widely used because of its advantages of non-contact and high recognition accuracy. However, the traditional RGB camera-based DMS has poor recognition accuracy under complex lighting conditions, while the IR-based DMS has a high cost. In order to improve the recognition accuracy of conventional RGB camera-based DMS under complicated illumination conditions, this paper proposes a lightweight low-illumination image enhancement network inspired by the Retinex theory. The lightweight aspect of the network structure is realized by introducing a pixel-wise adjustment function. In addition, the optimization bottleneck problem is solved by introducing the shortcut mechanism.

Harmony-with-traffic refers to the ability of autonomous vehicles to maximize the driving benefits such as comfort, efficiency, and energy consumption of themselves and the surrounding traffic during interactive driving under traffic rules. In the test of harmony-with-traffic, one or more background vehicles that can respond to the driving behavior of the vehicle under test are required. For this purpose, the functional requirements of car-following model for harmony-with-traffic evaluation are analyzed from the dimensions of test conditions, constraints, steady state and dynamic response. Based on them, an interactive car-following model (ICFM) is developed. In this model, the concept of equivalent distance is proposed to transfer lateral influence to longitudinal. The calculation methods of expected speed are designed according to the different car-following modes divided by interaction object, reaction distance and equivalent distance.

High-speed vehicles in low illumination environments severely blur the images used in object detectors, which poses a potential threat to object detector-based advanced driver assistance systems (ADAS) and autonomous driving systems. Augmenting the training images for object detectors is an efficient way to mitigate the threat from motion blur. However, little attention has been paid to the motion of the vehicle and the position of objects in the traffic scene, which limits the consistence between the resulting augmented images and traffic scenes. In this paper, we present a vehicle kinematics-based image augmentation algorithm by modeling and analyzing the traffic scenes to generate more realistic augmented images and achieve higher robustness improvement on object detectors against motion blur. Firstly, we propose a traffic scene model considering vehicle motion and the relationship between the vehicle and the object in the traffic scene.

The research and development of data-driven highly automated driving system components such as trajectory prediction, motion planning, driving test scenario generation, and safety validation all require large amounts of naturalistic vehicle trajectory data. Therefore, a variety of data collection methods have emerged to meet the growing demand. Among these, camera-equipped drones are gaining more and more attention because of their obvious advantages. Specifically, compared to others, drones have a wider field of bird's eye view, which is less likely to be blocked, and they could collect more complete and natural vehicle trajectory data. Besides, they are not easily observed by traffic participants and ensure that the human driver behavior data collected is realistic and natural. In this paper, we present a complete vehicle trajectory data extraction framework based on aerial videos. It consists of three parts: 1) objects detection, 2) data association, and 3) data cleaning.

Multiple object detection and tracking are central aspects of modeling the environment of autonomous vehicles. Lidar is a necessary component in the autonomous driving system. Without Lidar sensors, we will most probably not see fully self-driving cars become a reality. Lidar sensing gives us high-resolution data by sending out thousands of laser signals. In advanced driver assistance systems or automated driving systems, 3-D point clouds from lidar scans are typically used to measure physical surfaces. Lidar is a powerful sensor that you can use in challenging environments where other sensors might prove inadequate. Lidar can provide a complete 360-degree view of a scene. This paper designs Lidar based multi-target detection and tracking system based on the traditional point cloud processing method including down-sampling, denoising, segmentation, and clustering objects.

Multi-Target tracking is a central aspect of modeling the surrounding environment of autonomous vehicles. Automotive millimeter-wave radar is a necessary component in the autonomous driving system. One of the biggest advantages of radar is it measures the velocity directly. Another big advantage is that the radar is less influenced by environmental conditions. It can work day and night, in rainy or snowy conditions. In the expressway scenario, the forward-looking radar can generate multiple objects, to properly track the leading vehicle or neighbor-lane vehicle, a multi-target tracking algorithm is required. How to associate the track and the measurement or data association is an important question in a multi-target tracking system. This paper applies the nearest-neighbor method to solve the data association problem and uses an extended Kalman filter to update the state of the track.

Aiming at the problems of ineffective collision avoidance and vehicle instability in the process of vehicle emergency braking in road conditions with low adhesion and sudden change in adhesion coefficient, a stability-coordinated emergency braking and collision avoidance control system SEBCACS) is proposed. First, according to the motion of the ego vehicle and the target vehicle as well as the road adhesion conditions, a collision time model is proposed for evaluating the vehicle collision risk, and the expected deceleration required to avoid the collision is calculated. Then, the MPC method is used to calculate the yaw moment generated by the four-wheel braking force required to maintain vehicle stability according to the actual and reference yaw rate and side slip angle deviation. Then it is decided whether to implement additional yaw moment control according to the body stability evaluation results.

Motivated by applying artificial intelligence technology to the automobile industry, reinforcement learning is becoming more and more popular in the community of intelligent driving research. The reward function is one of the critical factors which affecting reinforcement learning. Its design principle is highly dependent on the features of the agent. The agent studied in this paper can do perception, decision-making, and motion-control, which aims to be the assistant or substitute for human driving in the latest future. Therefore, this paper analyzes the characteristics of excellent human driving behavior based on the six-layer model of driving scenarios and constructs it into a human knowledge graph. Furthermore, for highway pilot driving, the expert demo data is created, and the reward function is self-learned via inverse reinforcement learning. The reward function design method proposed in this paper has been verified in the Unity ML-Agent environment.

Since data processing methods could not completely eliminate the uncertainty of signals, it is a key issue for stable and robust decision-making for uncertainty tolerance of intelligent vehicles. In this paper, a decision-making for an Autonomous Emergency Braking (AEB) case considering the uncertainty of road adhesion coefficient estimation (RACE) is proposed. Firstly, the 3σ criterion is employed to classify the confidence in order to establish the decision-making mechanism considering the signal uncertainty of RACE. Secondly, the model for AEB with the uncertainty of the road adhesion coefficient estimated is designed based on the Seungwuk Moon model. Thirdly, a CCRs and CCRm scenario was designed to verify the feasibility in reference to the European New Car Assessment Programme (Euro NCAP) standard. Finally, the results of 10,000 cycles test illustrate that the proposed method is stable and could significantly improve the safety confidence both in the CCRs and CCRm scenarios.

Tracking vehicle trajectories is essential for autonomous vehicles and advanced driver-assistance systems to understand traffic environment and evaluate collision risk. In order to reduce the position deviation and fluctuation of tracking on-road vehicle by millimeter-wave radar (MMWR), an interactive multi-model Kalman filter (IMM-KF) tracking algorithm including data association and track management is proposed. In general, it is difficult to model the target vehicle accurately due to lack of vehicle kinematics parameters, like wheel base, uncertainty of driving behavior and limitation of sensor’s field of view. To handle the uncertainty problem, an interacting multiple model (IMM) approach using Kalman filters is employed to estimate multitarget’s states. Then the compensation of radar ego motion is achieved, since the original measurement is under the radar polar coordinate system.

Multi-sensor information fusion framework is the eyes for unmanned driving and Advanced Driver Assistance System (ADAS) to perceive the surrounding environment. In addition to the perception of the surrounding environment, real-time vehicle localization is also the key and difficult point of unmanned driving technology. The disappearance of high-precision GPS signal suddenly and defect of the lane line will bring much more difficult and dangerous for vehicle localization when the vehicle is on unmanned driving. In this paper, a road boundary feature extraction algorithm is proposed based on multi-sensor information fusion of automotive radar and vision to realize the auxiliary localization of vehicles. Firstly, we designed a 79GHz (78-81GHz) Ultra-Wide Band (UWB) millimeter-wave radar, which can obtain the point cloud information of road boundary features such as guardrail or green belt and so on.

Safety is the cornerstone for Advanced Driver Assistance Systems (ADAS) and Autonomous Driving Systems (ADS). To assess the safety of a traffic situation, it is essential to predict motion states of traffic participants in the future with mathematic models. Accurate vehicle trajectory prediction is an important prerequisite for reasonable traffic situation risk assessment and appropriate decision making. Vehicle trajectory prediction methods can be generally divided into motion model based methods and maneuver model based methods. Vehicle trajectory prediction based on motion models can be accurate and reliable only in the short term. While vehicle trajectory prediction based on maneuver models present more satisfactory performance in the long term, these maneuver models rely on machine learning methods. Abundant data should be collected to train the maneuver recognition model, which increases complexity and lowers real-time performance.

Automatic Lane Change (ALC) function is an important step to promote the currently popular Advanced Driver Assistance Systems (ADAS) within a single lane. The key issue for ALC is accurate steering angle and trajectory tracking during the lane changing process. In this paper, an MPC controller with a receding horizon is designed to track the desired trajectory. During the tracking process, other objectives such as safety and smoothness are considered. Considering of the practical mechanism and parameter uncertainties, an SMC controller is designed to track the target steering angle. For validation, a Hardware-in-the-Loop (HIL) experiment platform is established, and experiments of different control algorithms under different conditions are carried out successively. Comparisons of the experiment results of MPC+SMC and PID+SMC schemes indicate that both the trajectory error and the steering angle error of the former combination are smaller.

Lane Keeping Assistance System (LKAS) is a typical lateral driver assistance system with low acceptance. One of the main reasons is that fixed parameters cannot satisfy individual differences. So LKAS adaptive to driver characteristics needs to be designed. Driver Steering Override (DSO) process is an important process of LKAS. It happens when contradiction between driver’s intention and system behavior occurs. As feeling of overriding will affect the overall experience of using LKAS, the design of DSO characteristics is worthy of attention. This research provided an adaptive design scheme aiming at DSO characteristics for LKAS by building Driver Preference Model (DPM) based on simulator test data from preliminary experiments. The DPM was to represent the relationship between driver characteristics indices and driver preferred system characteristics indices. So that new drivers’ preference can be predicted by DPM based on their own daily driving data with LKAS switched off.

Automatic Parking Assist System (APAS) is an important part of Advanced Driver Assistance System (ADAS). It frees drivers from the burden of maneuvering a vehicle into a narrow parking space. This paper deals with the motion planning, a key issue of APAS, for vehicles in automatic parking. Planning module should guarantee the robustness to various initial postures and ensure that the vehicle is parked symmetrically in the center of the parking slot. However, current planning methods can’t meet both requirements well. To meet the aforementioned requirements, a method combining neural network and Monte-Carlo Tree Search (MCTS) is adopted in this work. From a driver’s perspective, different initial postures imply different parking strategies. In order to achieve the robustness to diverse initial postures, a natural idea is to train a model that can learn various strategies.

For the personification of automotive vehicle function performance under common traffic scenarios, analysis of human driver behavior is necessary. Based on China Field Operational Test (China-FOT) database of China Natural Driving Study project, this paper studies the driver's response in the common cut-in scenario. A total of 266 cut-in cases are selected by manual interception of driving recorder video. The relevant traffic environment characteristics are also extracted from video, including light conditions, road conditions, scale and lateral position of cut-in vehicle, etc. Dynamic information is decoded form CAN, such as speed, acceleration and so on. Then image processing results, such as relative speed and distance of cut-in and subject vehicles, are calculated. Statistical results based on above information show the response type and distribution of human driver: the behavior of keeping lane is 96.24%, in which the ratio of braking response is 51.13%.

Cut-in scenario is common in traffic and has potential collision risk. Human driver can detect other vehicle’s cut-in intention and take appropriate maneuvers to reduce collision risk. However, autonomous driving systems don’t have as good performance as human driver. Hence a deeper understanding on driving behavior is necessary. How to make decisions like human driver is an important problem for automated vehicles. In this paper, a method is proposed to classify the dangerous cut-in situations and normal ones. Dangerous cases were extracted automatically from naturalistic driving database using specific detection criteria. Among those cases, 70 valid dangerous cut-in cases were selected manually. The largest deceleration of subject vehicle is over 4 m/s2. Besides, 249 normal cut-in cases were extracted by going through video data of 2000km traveled distance. In normal driving cases, subject vehicle may brake or keep accelerating and the largest deceleration was less than 3 m/s2.

Cut-in scenarios are common and of potential risk in China but Advanced Driver Assistant System (ADAS) doesn’t work well under such scenarios. In order to improve the acceptance of ADAS, its reactions to Cut-in scenarios should meet driver’s driving habits and expectancy. Brake is considered as an express of risk and brake tendency in normal Cut-in situations needs more investigation. Under critical Cut-in scenarios, driver tends to brake hard to eliminate collision risk when cutting in vehicle right crossing lane. However, under less critical Cut-in scenarios, namely normal Cut-in scenarios, driver brakes in some cases and takes no brake maneuver in others. The time when driver initiated to brake was defined as key time. If driver had no brake maneuver, the time when cutting-in vehicle right crossed lane was defined as key time. This paper focuses on driver’s brake tendency at key time under normal Cut-in situations.

The camera-based advanced driver assistance systems (ADAS) like lane departure warning system (LDWS) and lane keeping assist (LKA) can make vehicles safer and driving easier. Lane detection is indispensable for these lane-based systems for achieving vehicle local localization and behavior prediction. Since the vision is vulnerable to the variable environment conditions such as bad weather, occlusions and illumination, the robustness is important. In this paper, a robust algorithm for detecting and tracking multiple lanes with arbitrary shape is proposed. We extend the previously lane detection and tracking process from the space domain to the temporal-spatial domain by using a more robust and general multi-lane model. First, new slice images containing temporal information are generated from image sequences. Instead of binarization process, we use a more general detector for extracting the lane marker candidates with prior knowledge to generate the binary slice image.

Driver behaviors provide abundant information and feedback for future Advanced Driver Assistance Systems (ADAS). Driver’s eye and head may present some typical movement patterns before executing driving maneuvers. It is possible to use driver’s head and eye movement information for predicting driver intention. Therefore, to determine the most important features related to driver intention has attracted widespread research interests. In this paper, a method to analyze the correlation between driver’s visual features and driver intention is proposed, aiming to determine the most representative features for driver intention prediction. Firstly, naturalistic driving experiment is conducted to collect driver’s videos during executing lane keeping and lane change maneuvers. Then, driver’s head and face visual features are extracted from those videos. By using boxplot and independent samples T-test, features which have significant correlation with driver intention are found.