Kwang-Ryul Baek

Implementation of the Ultrasonic Local Positioning System using Dual Frequencies and Codes

This paper presents real-time algorithm for an ultrasonic Local Positioning System(LPS). An ultrasonic LPS consists of 4 transmitters and n receivers. Each transmitter transmits an sequential ultrasonic signal to avoid interference of ultrasonic signal. This method is a noneffective application for a fast object. Because receiver detects four sequential transmissive ultrasonic signal and calculates a position. This paper proposes the method which 4 transmitters transmit simultaneous ultrasonic signal and each transmitter distinguished by frequencies and codes. And Auto-Correlation Function(ACF) method separates codes from an ultrasonic echo signal which is interference of each transmitter`s code. If the receiver uses only ACF method, it is difficult to implement real time application for increased computation. This paper implements LPS using dual frequencies and ACF method. Using dual frequencies reduces codes length. The reduced codes length save computation in ACF. To prove this algorithm by experiment, high performance DSP(digital signal processor) used. The result shows the performance of the designed system is good enough positioning.

Indoor Positioning System Using Magnetic Field Map Navigation and an Encoder System

In the indoor environment, variation of the magnetic field is caused by building structures, and magnetic field map navigation is based on this feature. In order to estimate position using this navigation, a three-axis magnetic field must be measured at every point to build a magnetic field map. After the magnetic field map is obtained, the position of the mobile robot can be estimated with a likelihood function whereby the measured magnetic field data and the magnetic field map are used. However, if only magnetic field map navigation is used, the estimated position can have large errors. In order to improve performance, we propose a particle filter system that integrates magnetic field map navigation and an encoder system. In this paper, multiple magnetic sensors and three magnetic field maps (a horizontal intensity map, a vertical intensity map, and a direction information map) are used to update the weights of particles. As a result, the proposed system estimates the position and orientation of a mobile robot more accurately than previous systems. Also, when the number of magnetic sensors increases, this paper shows that system performance improves. Finally, experiment results are shown from the proposed system that was implemented and evaluated.

Implementation of MOST/CAN network protocol

This paper describes the protocol of the in-vehicle network. Nowadays, many cars adopt in-vehicle network for lightweight and maintenance. CAN is popular in-vehicle network for control data, and MOST is network for multimedia data. But CAN and MOST does not integrate and work independently. This paper suggests integration of MOST and CAN. The two networks have different characteristics such as data frame structure, transfer speed, network/bus configuration, etc. So, we implement the MOST/CAN converter that integrates protocols of the two networks by transmitting the CAN data using the MOST asynchronous data channel. To verify the performance of the MOST/CAN converter, we adopt the queuing model and the SAE benchmark data.

Global thresholding algorithm based on boundary selection

A global image thresholding algorithm based on boundary selection is proposed for improving conventional histogram-based thresholding algorithms. An image is divided into blocks of a fixed size, and the pixel variance of each block is used for finding the boundary blocks that a bimodal histogram. Otsus threshold value of boundary block is used to classify object blocks and background blocks. The average values of these blocks are used to determine the optimal threshold value. Experimental results show that proposed algorithm improves on conventional thresholding methods for various types of images.

Position estimation using hyperbolic model for time difference of flight of ultrasonic

Three ultrasonic transmitters are in line at one side and transmit ultrasonic signals that contain different 9bit pseudo-random codes at the same time. The receiver measures time difference of flights (TDOFs) of transmitted signals by optimum correlation detection. According to TDOF, the possible position of the receiver is represented by hyperbolic curves. An intersection of the hyperbolic curves is the position of the receiver. The intersection is calculated by bisection method. The proposed system can be installed easily and the cost of installation is low. In the experimental results, the position errors become large when the receiver is at the position far from central transmitter and are correlated about (u, v) position.

Environment Adaptation using Evolutional Interactivity in a Swarm of Robots

In this paper we consider the multi-robot system that collects target objects spread in an unexplored environment. The robots cooperate each other to improve the capability and the efficiency. The robots attract or intimidate each other as behaviors of bacterial swarms or particles with electrical moments. The interactions would increase the working efficiency in some environments but it would decrease the efficiency in some other environments. Therefore, the system needs to adapt to the working environment by adjusting the strengths of the interactions. The strengths of the interactions are expressed as sets of gene codes that mean the weights of each kind of attracting or intimidating vectors. The proposed system adjusts the gene codes using evolutional strategy. The proposed approach has been validated by computer simulation. The results of this paper show that our inter-swarm interacting strategy and optimizing algorithm improves the working efficiency, adaptively to the characteristics of environments.

A data reorganization algorithm to improve transmission efficiency in CAN networks

CAN network is multi-master/multi-slave communication protocol and also supports single-master/multi-slave. In single-master/multi-slave mode, CAN master transmits control messages to each slave. If many slaves need small data from CAN master, CAN network spends a lot of time transmitting CAN overhead bits. So the transmission efficiency in CAN network goes down. In this paper, we propose an algorithm that increases the transmission efficiency by reorganizing the data mathematically.

Pedestrian Detection Algorithm using a Gabor Filter Bank

Abstract: A Gabor filter is a linear filter used for edge detectionas frequency and orientation representations of Gabor filters are similar to those of the human visual system. In this thesis, we propose a pedestrian detection algorithm using a Gabor filter bank. In order to extract the features of the pedestrian, we use various image processing algorithms and data structure algorithms. First, color image segmentation is performed to consider the information of the RGB color space. Second, histogram equalization is performed to enhance the brightness of the input images. Third, convolution is performed between a Gabor filter bank and the enhanced images. Fourth, statistical values are calculated by using the integral image (summed area table) method. The calculated statistical values are used for the feature matrix of the pedestrian area. To evaluate the proposed algorithm, the INRIA pedestrian database and SVM (Support Vector Machine) are used, and we compare the proposed algorithm and the HOG (Histogram of Oriented Gradient) pedestrian detector, presentlyreferred to as the methodology of pedestrian detection algorithm. The experimental results show that the proposed algorithm is more accurate compared to the HOG pedestrian detector. Keywords: pedestrian detection, Gabor filter, Gabor filter bank

Modified Otsu's method for indoor mobile robot tracking system

In vision-based tracking system, thresholding is one of the most important steps in image pre-processing. Thresholding algorithm has a strong influence on both accuracy and performance in object tracking. Thresholding algorithms are classified as global thresholding or local thresholding. In general, the computing power required for local thresholding algorithm is more than ten times that of global thresholding algorithm, so global thresholding algorithm is suitable for a real-time application. The Otsus method is the most famous global thresholding algorithm, however, it misclassifies object as background in some cases. To reduce the misclassification problems, we apply the modified Otsus method for indoor mobile robot tracking system. Experimental results show that applied algorithm improves the performance of thresholding results.

Position estimation using time difference of flight of the multi-coded ultrasonic

The environment where mobile robots move is modeled as 2-dimensional space. Three ultrasonic receivers are in a line. The transmitters attached to the mobile robots transmit ultrasonic signals which contain different 9bit PR (pseudo-random) codes. The receiver calculates TDOFs (Time Difference of Flight) of each transmitted signals by CCF (Cross Correlation Function). According to TDOFs, the possible position of the mobile robot is represented by hyperbolic curves. The position of the mobile robot is an intersection of hyperbolic curves. The intersection is calculated by bisection method. The external device becomes simple compared to traditional method. The proposed system is verified experimentally.