Seongkeun Park

A New Evolutionary Particle Filter for the Prevention of Sample Impoverishment

Particle filters perform the nonlinear estimation and have received much attention from many engineering fields over the past decade. Unfortunately, there are some cases in which most particles are concentrated prematurely at a wrong point, thereby losing diversity and causing the estimation to fail. In this paper, genetic algorithms (GAs) are incorporated into a particle filter to overcome this drawback of the filter. By using genetic operators, the premature convergence of the particles is avoided and the search region of particles enlarged. The GA-inspired proposal distribution is proposed and the corresponding importance weight is derived to approximate the given target distribution. Finally, a computer simulation is performed to show the effectiveness of the proposed method.

A new weighted approach to imbalanced data classification problem via support vector machine with quadratic cost function

In this paper, a new weighted approach on Lagrangian support vector machine for imbalanced data classification problem is proposed. The weight parameters are embedded in the Lagrangian SVM formulation. The training method for weighted Lagrangian SVM is presented and its convergence is proven. The weighted Lagrangian SVM classifier is tested and compared with some other SVMs using synthetic and real data to show its effectiveness and feasibility.

A neural network approach to target classification for active safety system using microwave radar

As a sensor in the active safety system of vehicles, the microwave radar (MWR) would be a good choice for the localization of the nearby targets but could be a bad choice for their classification or identification. In this paper, a target classification system using a 24GHz microwave radar sensor is proposed for the active safety system. The basic idea of this paper is that the pedestrians and the vehicles have different reflection characteristics for a microwave. A multilayer perceptron (MLP) neural network is employed to classify the targets and the probabilistic fusion is conduct over time to improve the classification accuracy. Some experiments are performed to show the validity of the proposed system.

Pedestrian/Vehicle Detection Using a 2.5-D Multi-Layer Laser Scanner

Laser scanners are widely used as the primary sensor for autonomous driving. When the commercialization of autonomous driving is considered, a 2.5-D multi-layer laser scanner is one of the best sensor options. In this paper, a new method is presented to detect pedestrians and vehicles using a 2.5-D multi-layer laser scanner. The proposed method consists of three steps: segmentation; feature extraction; and classification; this paper focuses on the last two steps. In feature extraction, new features for the multi-layer laser scanner are proposed to improve the classification performance. In classification, radial basis function additive kernel support vector machine is employed to reduce the computation time while maintaining the performance. The proposed method is implemented on a real vehicle, and its performance is tested in a real-world environment. The experiments indicate that the proposed method has good performance in many real-life situations.

A new state estimation method for chaotic signals: Map-particle filter method

Over the past few decades, research has been conducted regarding the applicability of chaotic behavior in consumer and industrial applications, including secure communication. In this paper, we develop a new particle filter termed as MAP-particle filter and apply it to a secure communication. The proposed particle filter estimates not only the chaotic state but also the secure message, thereby improving the secure communication performance. A computer simulation is conducted to demonstrate the effectiveness of our proposed method.

Camera and LIDAR Combined System for On-Road Vehicle Detection

In this paper, we design an on-road vehicle detection system based on the combination of a camera and a LIDAR system. In the proposed system, the candidate area is selected from the LIDAR data using a grouping algorithm. Then, the selected candidate area is scanned by an SVM to find an actual vehicle. The morphological edged images are used as features in a camera. The principal components of the edged images called eigencar are employed to train the SVM. We conducted experiments to show that the on-road vehicle detection system developed in this paper demonstrates about 80% accuracy and runs with 20 scans per second on LIDAR and 10 frames per second on camera.

Collision Risk Assessment for Pedestrians` Safety Using Neural Network

Abstract: This paper proposes a new collision risk assessment system for pedestrians’s safety. Monte Carlo Simulation (MCS) method is a one of the most popular method that rely on repeated random sampling to compute their result, and this method is also proper to get the results when it is unfeasible or impossible to compute an exact result. Nevertheless its advantages, it spends much time to calculate the result of some situation, we apply not only MCS but also Neural Networks in this problem. By Monte carlo method, we make some sample data for input of neural networks and by using this data, neural networks can be trained for computing collision probability of whole area where can be measured by sensors. By using this trained networks, we can estimate the collision probability at each positions and velocities with high speed and low error rate. Computer simulations will be shown the validity of our proposed method. Keywords: intelligent vehicle, monte carlo, neural networks, collision risk, monte carlo simulation

On-road Vehicle Tracking using Laser Scanner with Multiple Hypothesis Assumption

Active safety vehicle devices are getting more attention recently. To prevent traffic accidents, the environment in front and even around the vehicle must be checked and monitored. In the present applications, mainly camera and radar based systems are used as sensing devices. Laser scanner, one of the sensing devices, has the advantage of obtaining accurate measurement of the distance and the geometric information about the objects in the field of view of the laser scanner. However. there is a problem that detecting object occluded by a foreground one is difficult. In this paper, criterions are proposed to manage this problem. Simulation is conducted by vehicle mounted the laser scanner and multiple-hypothesis algorithm tracks the candidate objects. We compare the running times as multi-hypothesis algorithm parameter varies.

Dual margin approach on a Lagrangian support vector machine

In this paper, we propose a new support vector machine (SVM) called dual margin Lagrangian support vectors machine (DMLSVM). Unlike other SVMs which use only support vectors to determine the separating hyperplanes, DMLSVM utilizes all the available training data for training the classifier, thus producing robust performance. The training data are weighted differently depending on whether they are in a marginal region or surplus region. For fast training, DMLSVM borrows its training algorithm from Lagrangian SVM (LSVM) and tailors the algorithm to its formulation. The convergence of our training method is rigorously proven and its validity is tested on a synthetic test set and UCI dataset. The proposed method can be used in a variety of applications such as a recommender systems for web contents of IPTV services.

Multiple data association and tracking using millimeter wave radar

In this paper, in order to support convenience and safety to drivers, we track out the moving vehicle. A target makes several sensor measurements and we have to cluster these several measurements. In order to cluster multiple sensor measurements, we use geometric information of sensor measurements. 24 GHz millimeter wave sensor is used in our proposed method. Real world experiments show the superiority our proposed method.