Daniel Alexander Meissner

Cooperative multi sensor network for traffic safety applications at intersections

To significantly reduce injury and fatal accidents smart intersections equipped with sensors and communication infrastructure have been proposed. In this publication a novel multi sensor network to perceive the intersection environment is presented. Based on an intensive analysis of accident scenarios in Germany the system was designed to address 75 % of all severe and lethal accidents. 14 laserscanners, 10 cameras, signal phase tapping and an I2V communication unit have been installed at a public intersection in Aschaffenburg, Germany. By using computer based field of view modelling the sensor positions are carefully selected to avoid occlusions. Thus, the infrastructure perception system provides a birds eye view. Our experiments show that spatial and temporal alignment of sensor data is achieved. We also demonstrate that a part of the sensor network, a calibrated stereo system, allows 3D coordinates in the field of view region of the cameras to be determined with an accuracy of 30 mm.

Road user tracking at intersections using a multiple-model PHD filter

A major aim of the joint project Ko-PER is the mitigation of fatal accidents at urban intersections. Therefore several test intersections have been equipped with multiple laser range finders to recognize and track road users. Besides a high traffic density the variety of road users is challenging. In this contribution a multiple-model (MM) probability hypothesis density filter with a track representation extended by class probabilities is proposed. The approach enables tracking of road users with appropriate motion models using a single MM filter. Due to the estimation of the class probabilities an adaption of the transition probabilities between the models is possible. The performance of the road user tracking is evaluated using real world data.

Intersection-Based Road User Tracking Using a Classifying Multiple-Model PHD Filter

The number of fatal accidents involving pedestrians and bikers at urban intersections is still increasing. Therefore, an intersection-based perception system provides a dynamic model of the intersection scene to the vehicles. Based on that, the intersection perception facilitates to discriminate occlusions which is expected to significantly reduce the number of accidents at intersections. Therefore this contribution presents a general purpose multi-sensor tracking algorithm, the classifying multiple-model probability hypothesis density (CMMPHD) filter, which facilitates the tracking and classification of relevant objects using a single filter. Due to the different motion characteristics, a multiple-model approach is required to obtain accurate state estimates and persistent tracks for all types of objects. Additionally, an extension of the PHD filter to handle contradictory measurements of different sensor types based on the Dempster-Shafer theory of evidence is proposed. The performance of tracking and classification is evaluated using real world sensor data of a public intersection.

Multi-object tracking at intersections using the cardinalized probability hypothesis density filter

A large percentage of accidents with body injuries in urban areas occur at intersections. Thus, improving safety at intersections using infrastructure based perception systems is desirable. In order to recognize and track the moving objects, a network of laserscanners is used to observe the intersection. In this contribution, a robust object recognition algorithm for vehicles and pedestrians is proposed. Further, the Cardinalized Probability Hypothesis Density with integrated estimation of the clutter density is applied to track vehicles and pedestrians at an intersection. The performance of the system is evaluated using real world sensor data of an intersection.

Real-time detection and tracking of pedestrians at intersections using a network of laserscanners

Accident analysis shows that the majority of accidents with body injuries occur in urban areas and more than 50 percent of those urban accidents happen at intersections. Due to that a major aim of the Ko-PER project, which is part of research initiative Ko-FAS, is to improve safety at intersections by infrastructure based perception. To recognize and track the moving objects, a network of laserscanner sensors observes the intersection and provides a 3D profile of the current scene. By means of the 3D measurements a robust and adaptive Gaussian mixture background model is trained to segment the measurements of dynamic objects and static objects. After the segmentation, the foreground points of each sensor are clustered based on the density of the point clouds and finally pedestrians are classified using dimension features. This paper focuses on tracking of pedestrians, which are the most vulnerable road users. In order to be able to integrate dependencies between the states of the pedestrians, a random finite set particle filter is used to track the pedestrians. The performance of the laserscanner based tracking system is shown and evaluated with measurements from the Ko-PER test intersection at Conti-Safety-Park. Therefore, the optimal subpattern assignment (OSPA) metric is used to evaluate the object recognition and tracking system.

Simulation and calibration of infrastructure based laser scanner networks at intersections

Accident analysis shows that intersections are a focal point for accidents in urban areas. Due to that, traffic monitoring at intersections has attached much attention. A major part of the Ko-PER project, which is part of the research initiative Ko-FAS, promoted by the Federal Ministry of Economics and Technology of Germany, is the infrastructure based perception of all dynamic objects inside an intersection. In this project, a novel system of detecting and tracking objects inside intersections using multiple 4-layer laser scanners is proposed. To reduce occlusions and maximize the observed area the sensors are mounted high over ground-level to achieve a birds eye view of the scene. One difference between laser scanners and video cameras is the difficulty to identify the area of the street, which can be observed by laser sensors, especially when the monitored area is plain like roads. Therefore, a realistic 3D simulation of the urban intersection and the laser range scanners was implemented. Based on this simulation, a method to calibrate the sensors was developed. The technique is easy to use and due to the 3D model of the intersection we were able to verify the proposed calibration tool.

Vehicle detection and tracking at intersections by fusing multiple camera views

Intersections are challenging locations for drivers. Complex situations are common due to the variety of road users and intersection layouts. This contribution describes a real time method for detecting and tracking vehicles at intersections using images captured by a static camera network. After background subtraction, the foreground segments are projected on a common fusion map. Using this fusion map, the pose, width, and height of the vehicles can be determined. After that, the detected objects are tracked by a Gaussian-Mixture approximation of the Probability Hypothesis Density filter. Results of the intersection perception can further be communicated to equipped vehicles by wireless communication.

The Ko-PER Intersection Laserscanner and Video Dataset

Public intersections are due to their complexity challenging locations for drivers. Therefore the german joint project Ko-PER - which is part of the project initiative Ko-FAS has equipped a public intersection with several laserscanners and video cameras to generate a comprehensive dynamic model of the ongoing traffic. Results of the intersection perception can be communicated to equipped vehicles by wireless communication. This contribution wants to share a dataset of the Ko-PER intersection to the research community for further research in the field of multi-object detection and tracking. Therefor the dataset consists of sensor data from the laserscanners network and cameras as well as reference data and object labels. With that dataset, we aim to stimulate further research in this area.

Divergence detectors for the δ-generalized labeled multi-Bernoulli filter

In single-target tracking, divergence detectors like the normalized innovation squared (NIS) are used to detect if the assumed motion or measurement models deviate too much from the actual behavior of the tracked target or the sensor. A generalization of the divergence detectors to random finite set based multi-object tracking algorithms is possible and results in the multi-target generalized NIS (MGNIS). In this contribution the MGNIS for the δ-generalized labeled multi-Bernoulli filter is derived. Further, an approximate multi-target NIS (AMNIS) is proposed which facilitates easier interpretation of the results. The MGNIS and the AMNIS are compared to the well-known optimal subpattern assignment (OSPA) metric using simulated data with different clutter rates.