Yangzhou Chen

Hybrid petri net modeling of traffic flow and signal control

Traffic control systems of signalized intersections are naturally hybrid systems, in which vehicle flow behavior can be described by a time-driven model and the traffic light dynamics are modeled as a discrete event system. In this paper, we use hybrid Petri nets (HPNs) to specify traffic and traffic control at an intersection and use a second-order macroscopic model to model the motion of vehicles in a road stretch between two successive intersections. The traffic control system of an arterial street is thus modeled as a composition of individual intersection models and road stretch models. Such a structure is suitable to perform traffic optimization by changing the length of the offset between two adjacent intersections with the common cycle.

Driver fatigue detection based on mouth information

Drivers fatigue detection has been realized based on drivers mouth geometrical features. It can give some information when driver is fatigue. For better speed and reliability, a new method which was based on combined Adaboost algorithm and particle filter was applied. Then the mouth verification was applied according to prior knowledge. Drivers state was judged by the geometrical features in a period of time. As a result, the detection accuracy was improved. This method can meet the requirement of drivers fatigue detection.

Vehicle queue detection based on morphological edge

In an Intelligent Traffic Systems(ITS), queue parameters at the intersections are required in many applications, such as accident monitoring and the controlling of traffic light. In this paper we use morphological edge detection to deal with the traffic image and propose a flexible window method to detect the length of the queue, which adapts to the characteristic and requirement of queue detection. The method is better than background differentiation and frame differentiation in vehicle detection and queue detection. Experimental results show that method is accurate and can run in real-time under different weather, illumination and occlusions.

Weather Condition Recognition Based on Feature Extraction and K-NN

Most of vision based transport parameter detection algorithms are designed to be executed in good-natured weather conditions. However, limited visibility in rain or fog strongly influences detection results. To improve machine vision in adverse weather situations, a reliable weather conditions detection system is necessary as a ground base. In this article, a novel algorithm for weather condition automatic recognition is presented. This proposed system is able to distinguish between multiple weather situations based on the classification of single monocular color images without any additional assumptions or prior knowledge. Homogenous area is extracted form top to bottom in scene image. Inflection point information which implies visibility distance will be taken as a character feature for current weather recognition. Another four features: power spectral slope, edge gradient energy, contrast, saturation, and image noisy which descript image definition are extracted also. Our proposed image descriptor clearly outperforms existing descriptors for the task. Experimental results on real traffic images are characterized by high accuracy, efficiency, and versatility with respect to driver assistance systems.

Longitudinal control of intelligent vehicle based on hybrid automata model

Based on the analysis of longitudinal control scenarios, a longitudinal control model of the intelligent vehicle in virtue of hybrid automata is built. Driving situations are divided into cruise control, speed following, inter-vehicle distance adjustment modes. Three corresponding control strategies are proposed depending on the driving modes. Then an algorithm of the longitudinal controller is designed. Not only is the following control of the target vehicle in the same lane achieved, but also, when the target vehicle changed, the following control of the new target vehicle is achieved through the control strategies of inter-vehicle distance adjustment. Simulation results show that the designed strategy can achieve a variety of scenarios in both high-speed driving and low-speed stop-and-go situations.

Traffic flow characteristic analysis at intersections from multi-layer spectral clustering of motion patterns using raw vehicle trajectory

Road intersections are important components of urban road system. It is the traffic flow characteristics representing the current traffic situation that provide a basis for the planning, designing and management of intersections. In this paper, we constructed an automatic processing framework on traffic flow characteristics analysis and understanding the traffic state at the urban road intersections, based on the collected raw vehicle motion trajectories. Our proposed method is basically attributed to identifying distinct vehicle motion patterns at intersections hierarchically using raw trajectory. Firstly, the fundamental assumption in traditional approaches that the trajectory set of high quality are readily available after manual rectification is not taken for granted any more. And by fully analyzing the local characteristics of trajectories, we figure out and explain various patterns behind traffic flow as well as yielded higher accuracy in motion trajectory clustering under the multi-layer spectral clustering method. At last, coupling the analyzing results with the surrounding characteristics of the intersection, the examples computing traffic flow features and predicting vehicle activity illustrates the potential of applying vehicle trajectories to traffic study, which are all suggested by experimental results.

Path planning based on Constrained Delaunay Triangulation

This paper proposes a path planning algorithm for determining an optimal path with respect to the costs of a dual graph on the Constrained Delaunay Triangulation (CDT) of an environment. The advantages of using triangles for environment expression are: less data storage required, available mature triangulation methods and consistent with a potential motion planning framework. First we represent the polygon environment as a planar straight line graph (PSLG) described as a collection of vertices and segments, and then we adopt the CDT to partition the environment into triangles. Then on this CDT of the environment, a dual graph is constructed following the target attractive principle in order to avoid the nonoptimal paths caused by the different geometric size of the triangles. Correspondingly, a path planning algorithm via A* search algorithm finds an optimal path on the real-time building dual graph. In addition, completeness and optimization analysis of the algorithm is given. The simulation results demonstrate the effectiveness and optimization of the algorithm.

Traffic meteorological visibility estimation based on homogenous area extraction

A novel technique for estimating traffic visibility based on homogenous area extraction is presented in this paper. Focusing on the problem of detecting daytime fog and estimating visibility, this proposed algorithm adopts a measure defined by the International Commission on Illumination CIE as the distance beyond which a black object of an appropriate dimension is perceived with a contrast of less than 5%. The key step for the algorithm is homogenous area extraction including sky and road selection from the top to the bottom of traffic image. Traffic meteorological visibility will be known as long as the inflection points location of this area is found. Camera parameter calibration should be done by vanishing points calculation based on painted lines detection. At last, calibration results, inflection point and traffic meteorological visibility for three different traffic scenes will be given. Comparison with real data obtained manually verifies the effectiveness of the algorithm.

Weather identifying system based on vehicle video image

Aim at the influence of the weather factor upon the road vehicle, this paper puts forward a kind of a weather identifying system vehicle video image. The system needs to initialize an original background image which is the brightest on the average in the first period of time. The background image is renewed not only when the bright degree of next frame image is higher than it, but also the weather is detected sun day at the same time. Get a background image of the current video sequence and let it minus with it to obtain the identifying image. At last do texture analysis to it to realize the weather automatically identifying. This paper mainly identifies the weather in a sun day, cloudy day and snow day, and after a lot of experiments the method are proved to be effective, and it can identify the weather condition of the current vehicle video accurately.

Camshift-Based Real-Time Multiple Vehicle Tracking for Visual Traffic Surveillance

This paper presents methods for vision-based detection and tracking of vehicles in monocular image sequences of traffic scenes recorded by a stationary camera. The goal of this research is to develop suitable methods for automatic visual traffic surveillance to perform detection, tracking and traffic parameter estimation of multiple vehicles in real time as well as tackle environment illumination changes and vehicle occlusion. Each of detected vehicles is assigned a camshift tracker which can quickly and exactly track object with different size and shape. Experimental results from traffic scenes demonstrate the effectiveness and robustness of the methods.