Ieee Its Society Newsletter

s of forthcoming papers on IEEE Transactions on ITS PEDESTRIAN SAFETY ANALYSIS IN MIXED TRAFFIC CONDITION USING VIDEO DATA ZHANG, YINGYING; YAO, DANYA (DANIEL); QIU, TONY; PENG, LIHUI; ZHANG, YI With dramatic development of image processing technology, a growing number of traffic flow detection and analysis have been conducted by using video data. Time to Collision (TTC) and Post-Encroachment Time (PET) are two major parameters to indicate the potential collision’s severity and to capture imminent vehicular accident. However, microlevel pedestrian involved collisions are less studied because they are hard to observe or record. This paper tries to extract the traffic object locations from video data, define the Time Difference to Collision (TDTC) parameter as a variation from TTC and PET to fit the pedestrian involved potential collisions/conflicts, analyze the interaction behavior between pedestrian and vehicle, and validate the TDTC parameter in indicating pedestrian safety performance by using 100 groups of interaction data. The result shows that the interaction cases with larger TDTC values are safer, while the cases with continuously closer to zero TDTC values are more dangerous. About 80% of the cases classified by the TDTC parameter have the same result with the independent observation; if TDTC is combined with vehicle speed, the classification result can be improved. More mixed traffic scenes will be conducted based on this research in the future. DEVELOPMENT OF A SMALL-SCALE RESEARCH PLATFORM FOR INTELLIGENT TRANSPORTATION SYSTEMS SHENG, WEIHUA; LA, HUNG; LIM, RONNY; DU, JIANHAO; ZHANG, SIJIAN; YAN, GANGFENG In this paper, we propose and develop a small-scale research platform for intelligent transportation systems. Our platform has four main parts: an arena; an indoor localization system; automated radio controlled (RC) cars; and roadside monitoring facilities. First, to mimic traffic environments we build an arena with a wooden floor, mock buildings and streets. Second, to facilitate feedback control for trajectory following, an indoor localization system is set up to track the RC cars. Third, both autonomous driving RC cars and human driving RC cars are developed, based on an automated RC car design. The automated RC cars can receive control signals from a computer through an Xbee RF module and control the front and rear wheels through motors. A new control algorithm is developed to allow the RC cars to track predefined trajectories. Finally, we implement an example of roadside monitoring which uses a fisheye camera associated with advanced video processing for image segmentation, object identification and tracking. Experiments are performed to demonstrate the effectiveness of the designed platform. We also discuss possible ITS research problems which can be studied in this testbed. Vol.14, No.4 October 2012 29 A TASK ASSIGNMENT ALGORITHM FOR MULTIPLE AERIAL VEHICLES TO ATTACK TARGETS WITH DYNAMIC VALUE ZHONG, LIU; LUO, QUAN; WEN, DING; QIAO, SHIDONG; SHI, JIANMAI; ZHANG, WEIMING A good task assignment is an important guarantee to achieve great combat effectiveness. This paper investigates the task assignment problem where the value of the targets is time-changing in battlefield, and presents a solution approach which is a combination of two algorithms: the multi-destination route planning algorithm based on dynamic programming and the multi-subgroup ant colony algorithm (MSACO). The two algorithms coordinately solve the task assignment problem. The route planning algorithm can obtain available routes between any two targets and provide reasonable routing information for MSACO. Then ant colony algorithm is applied to solve the task assignment problem. To solve the task assignment problem in the battlefield environment, several key technologies are introduced to improve the traditional ant colony algorithm, which include subgroup selection strategy, dynamic candidate aggregate policy, state transferring policy and information element updating mechanism. Simulations results show that the proposed approach can produce reasonable and available plan for all the test cases in short computational time. DISTRIBUTED CLASSIFICATION OF TRAFFIC ANOMALIES USING MICROSCOPIC TRAFFIC VARIABLES THAJCHAYAPONG, SUTTIPONG; GARCIA TREVIÑO, EDGAR; BARRIA, JAVIER This paper proposes a novel anomaly classification algorithm that can be deployed in a distributed manner and utilizes microscopic traffic variables shared by neighbouring vehicles to detect and classify traffic anomalies under different traffic conditions. The algorithm which incorporates multi-resolution concepts is based on the likelihood estimation of a neural network output and a bisection-based decision threshold. We show that when applied to real-world traffic scenarios, the proposed algorithm can detect all the traffic anomalies of the reference test data set; this represents a significant improvement over our previously proposed algorithm. We also show that the proposed algorithm can effectively detect and classify traffic anomalies even when i) the microscopic traffic variables are available from only a fraction of the vehicle population and ii) some microscopic traffic variables are lost due to degradation in V2V and/or V2I communications. SPATIO-TEMPORAL TRAFFIC SCENE MODELING FOR OBJECT MOTION DETECTION HAO, JIUYUE; LI, CHAO; KIM, ZUWHAN; XIONG, ZHANG Moving object detection is an important component of a traffic surveillance system. Usual background subtraction approaches often perform poorly on a long outdoor traffic video due to vehicles waiting at an intersection and gradual changes of illumination and background shadow position. We present a fast and robust background subtraction algorithm based on unified spatio-temporal background and foreground modeling. The correlation between neighboring pixels provides high levels of detection accuracy in the dynamic background scene. Our Bayesian fusion method, which establishes the traffic scene model, combines both background and foreground models, and considers prior probabilities to adapt changes of background in each frame. We explicitly model both temporal and spatial information based the Kernel Density Estimation (KDE) formulation for background modeling. Then, we use a Gaussian formulation to describe spatial correlation of moving objects for foreground modeling. In the Updating Step, a fusion background frame is generated and reasonable updating rates are also proposed for the traffic scene. The Vol.14, No.4 October 2012 30 experimental results show that the proposed method outperforms the previous work with less computation, and is better suited for the traffic scenes. VEHICLE POSITIONING USING GSM AND CASCADE-CONNECTED ANN STRUCTURES BORENOVIC, MILOS; NESKOVIC, ALEKSANDAR; NESKOVIC, NATASA Procuring the location information for intelligent transportation systems is a popular topic amongst researchers. This paper investigates the vehicle location algorithm based upon the received signal strengths from the available GSM networks. The performances of positioning models, consisted of cascade-connected Artificial Neural Network (ANN) multilayer feedforward structures, employing space-partitioning principle, are compared to the single ANN multilayer feedforward model in terms of accuracy, number of subspaces and other positioning relevant parameters. Cascadeconnected ANN structures make use of the fact that a vehicle can be found only in a subspace of the entire environment (roads) to improve the positioning accuracy. The best performing cascade-connected ANN structure achieved an average error of 26m, and a median error of less than 5m, which is accurate enough for most of the vehicle location services. Using the same received signal strength database obtained by measurements it was shown that the proposed model outperforms kNN and EKF trained ANN positioning algorithms. Moreover, the presented ANN structures replace not only the positioning algorithms, but also the overloaded map-matching process. CLASSIFICATION AND COUNTING OF COMPOSITE OBJECTS IN TRAFFIC SCENES USING GLOBAL AND LOCAL IMAGE ANALYSIS SOMASUNDARAM, GURUPRASAD; SIVALINGAM, RAVISHANKAR; MORELLAS, VASSILIOS; PAPANIKOLOPOULOS, NIKOLAOS Object recognition algorithms often focus on determining the class of a detected object in a scene. There are usually two significant phases involved in object recognition. The first phase is the object representation phase in which the most suitable features that provide the best discriminative power under constraints such as lighting, resolution, scale and view variations are chosen to describe the objects. The second phase is to use this representation space to develop models for each object class using discriminative classifiers. In this paper we focus on composite objects, i.e. objects which have two or more simpler classes interconnected in a complicated manner. A classic example of such a scenario is a bicyclist. A bicyclist consists of a bicycle and a human riding it. When we are faced with the task of classifying bicyclists and pedestrians, it is counter-intuitive and often hard to come up with a discriminative classifier to distinguish the two classes. We explore global image analysis based on bag of visual words compare the results with local image analysis in which we attempt to distinguish the individual parts of the composite object. We also propose a unified Naive Bayes framework as well as a combined histogram feature method for combining the individual classifiers for enhanced performance. MODEL-INDEPENDENT ADAPTIVE FAULT-TOLERANT OUTPUT TRACKING CONTROL OF 4WS4WD ROAD VEHICLES LI, DANYONG; SONG, YONGDUAN; HUANG, DONG; CHEN, HENAN This paper investigates the path tracking control problem of four-wheel-steering and four-wheel-driving (4WS4WD) road vehicles. Of particular interest is the development of an adaptive and fault-tolerant tracking control scheme Vol.14, No.4 October 2012 31 capable of compensating vehicle uncertain dynamics/disturbances as well as actuation failures simultaneously. Control algorithms are derived without requiring detail system dynamic information. The control scheme is shown to be effective in coping with unexpected actuation faults without the need for analytically estimating bound on actuator failure variables. The proposed method is validated and demonstrated through its application to a wheeled vehicle with four steering wheels and four driving wheels, where high precision path tracking is achieved in the face of steering faults. SHORT-RUN ROUTE DIVERSION: AN EMPIRICAL INVESTIGATION INTO VARIABLE MESSAGE SIGN DESIGN AND POLICY EXPERIMENTS JINDAHRA, PAVITRA; CHOOCHARUKUL, KASEM Variable message signs (VMS) can convey several traffic and roadway information to motorists. Using empirical state preference (SP) data from road users in Bangkok, we demonstrate that short-run route diversion can be estimated and forecasted based on different VMS message content attributes via mixed logit and logit models in which the motorist’s stated route diversion is the dependent variable. The findings reveal that different message contents lead to different levels of route-changing propensity. Route diversion in Bangkok is likely when a VMS displays a suggested route and qualitative information. The framing effect on route choice decision explains the finding of qualitative delay information preference to its quantitative counterpart. To determine the policy implications, we further investigate the developed models by estimating changes in the probability of the stated route choice due to changes in the message content. Three VMS message policy experiments are conducted using the model: enforcing quantitative delay content, enforcing qualitative delay content, and enforcing suggested route content. The results show that qualitative delay information and suggested route reduce the ambiguity of the message quality. The optimal VMS designs for short-run traffic management to encourage/discourage route diversion are discussed. A FIXED SENSOR-BASED INTERSECTION COLLISION WARNING SYSTEM IN VULNERABLE LINEOF-SIGHT AND/OR TRAFFIC VIOLATION PRONE ENVIRONMENT JANG, JEONGAH; CHOI, KEECHOO; CHO, HANBYEONG This paper proposes a Cooperative Intersection Collision Warning System (CICWS) model which uses fixed traffic sensors to provide warning information to drivers at unsignalized intersections in a real-time manner. The CICWS model is useful for vulnerable line-of-sight and/or traffic violation prone environment since it determines whether the situation is really dangerous one or not. More specifically, the model is for unsignalized intersections without STOP/YIELD signs where drivers don’t tend to stop. The situation forecast model uses vehicle location, speed, and time data obtained from multiple sensors located at intersection approaches together with obstacle position and sight distance relationship. More specifically, the model has a real-time sight-distance triangle module and a collision-time prediction module. Using a micro traffic simulator called VISSIM, the validation and evaluation of the model are performed based on different scenarios with different parameters, like inflow volume, locations of traffic sensor, design speed, and obstacle placement. The results show that the model successfully forecasts dangerous situations up to 94.3%, which may imply the deployment of the model in such an environment where vehicle to infrastructure or vehicle to vehicle communication are possible. Some limitations and future research agenda have also been discussed. Vol.14, No.4 October 2012 32 ACCURATE GLOBAL LOCALIZATION USING VISUAL ODOMETRY AND DIGITAL MAPS ON URBAN ENVIRONMENTS PARRA ALONSO, IGNACIO; FERNANDEZ-LLORCA, DAVID; GAVILÁN, MIGUEL; ALVAREZ, SERGIO; GARCÍAGARRIDO, MIGUEL ANGEL; VLACIC, LJUBO; SOTELO VÁZQUEZ, MIGUEL ÁNGEL In the last years, Advanced Driver Assistance Systems (ADAS) have become a key element in the research and development of intelligent transportation systems and particularly, of intelligent vehicles. Many of these systems require accurate global localisation information, which has been traditionally performed by GPS’s, despite its wellknown failings, particularly in urban environments. Different solutions have been attempted to bridge the gaps of GPS positioning errors, but they usually require additional expensive sensors. Vision-based algorithms have proved to be capable of tracking the position of a vehicle over long distances using only a sequence of images as input and with no prior knowledge of the environment. This paper describes a full solution to the estimation of the global position of a vehicle in a digital road map by means of visual information alone. Our solution is based on a stereo platform used to estimate the motion trajectory of the ego vehicle and a map matching algorithm which will correct the cumulative errors of the vision-based motion information and estimate the global position of the vehicle in a digital road map. We demonstrate our system in large scale urban experiments reaching high accuracy in the estimation of the global position and allowing for longer GPS blackouts thanks to both the high accuracy of our visual odometry estimation and the correction of the cumulative error of the map matching algorithm. Typically challenging situations in urban environments such as non-static objects or illumination exceeding the dynamic range of the cameras are shown and discussed. RELIABLE CLASSIFICATION OF VEHICLE TYPES BASED ON CASCADE CLASSIFIER ENSEMBLES