• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.9
• /
• pp.369-374
• /
• 2014

Path generation methods are required for safe and efficient driving in unmanned autonomous vehicles. There are two kinds of paths: global and local. A global path consists of all the way points including the source and the destination. A local path is the trajectory that a vehicle needs to follow from a way point to the next in the global path. In this paper, we propose a novel method for local path generation through machine learning, with an effective curve function used for initializing the trajectory. First, reinforcement learning is applied to a set of candidate paths to produce the best trajectory with maximal reward. Then the optimal steering angle with respect to the trajectory is determined by training an artificial neural network. Our method outperformed existing approaches and successfully found quality paths in various experimental settings, including the cases with obstacles.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.8 no.6
• /
• pp.23-35
• /
• 2009

This study proposes a methodology for estimating representative section travel times using individual vehicle travel information under the ubiquitous transportation environment (UTE). A novel approach is to substantialize a concept of dynamic node-links in processing trajectory data. Also, grouping vehicles was conducted to obtain more reliable travel times representing characteristics of individual vehicle travels. Since the UTE allows us to obtain higher accuracy of vehicle positions, travel times for each lane can be estimated based on the proposed methodology. Evaluation results show that less than 10% of mean absolute percentage error was achievable with 20% of probe vehicle rate. It is expected that outcome of this study is useful for providing more accurate and reliable traffic information services.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.945-951
• /
• 2022

The major objective of this research is to evaluate performance of improved intersections for response time to emergency vehicle preemption. Smart technologies have been introduced to civil infrastructure systems for resilient communities. The technologies need to evaluate their effectiveness and feasibility to confirm their introduction. This research focuses on the performance of emergency vehicle preemption, represented by response time, when smart intersections are introduced in a community. The response time is determined by not only intersections but also a number of factors such as traffic, distance, road conditions, and incident types. However, the evaluation of emergency response has often ignored factors related to emergency vehicle routes. In this respect, this research synthetically analyzes geospatial and incident data using each route of emergency vehicle and conducts before-and-after evaluations. The changes in performance are analyzed by the impact of smart intersections on response time through Bayesian regression models. The result provides measures of the project's performance. This study will contribute to the body of knowledge on modeling the impacts of technology application and integrating heterogeneous data sets. It will provide a way to confirm and prove the effectiveness of introducing smart technologies to our communities.

• Proceedings of the KOR-KST Conference
• /
• 2007.05a
• /
• pp.179-187
• /
• 2007

• International Journal of Automotive Technology
• /
• v.8 no.5
• /
• pp.583-591
• /
• 2007

This paper describes the development of the $H_{\infty}$ lateral control system for an autonomous ground vehicle operating a limited area using the RTK-DGPS(Real Time Kinematic-Differential Global Positioning System). Before engaging in autonomous driving, map data are acquired by the RTK-DGPS and used to construct a reference trajectory. The navigation system contains the map data and computes the reference yaw angle of the vehicle using two consecutive position values. The yaw angle of the vehicle is controlled by the $H_{\infty}$ controller. A prototype of the autonomous vehicle by the navigation method has been developed, and the performance of the vehicle has been evaluated by experiment. The experimental results show that the $H_{\infty}$ controller and the RTK-DGPS based navigation system can sufficiently track the map at low speed. We expect that this navigation system can be made more accurate by incorporating additional sensors.

• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.12
• /
• pp.1132-1138
• /
• 2012

The MPC (Model Predictive Control) is one of the techniques that can be used to control an unmanned vehicle. It predicts the future vehicle trajectory using the dynamic characteristic of the vehicle and generate the control value to track the reference path. If some obstacles are detected on the reference paths, the MPC can generate control value to avoid the obstacles imposing the inequality constraints on the MPC cost function. In this paper, we propose an obstacle modeling algorithm for MPC with inequality constraints for obstacle avoidance and a method to set selective constraint on the MPC for stable obstacle avoidance. Simulations with the field test data show successful obstacle avoidance and way point tracking performance.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.62-65
• /
• 2018

This paper comes out with the study on sensing data collection strategy in a Software-Defined Mobile Edge vehicular networking. The two cooperative data dissemination are Direct Vehicular cloud mode and edge cell trajectory prediction decision mode. In direct vehicular cloud, the vehicle observe its neighboring vehicles and sets up vehicular cloud for cooperative sensing data collection, the data collection output can be transmitted from vehicles participating in the cooperative sensing data collection computation to the vehicle on which the sensing data collection request originate through V2V communication. The vehicle on which computation originate will reassemble the computation out-put and send to the closest RSU. The SDMEVN (Software Defined Mobile Edge Vehicular Network) Controller determines how much effort the sensing data collection request requires and calculates the number of RSUs required to support coverage of one RSU to the other. We set up a simulation scenario based on realistic traffic and communication features and demonstrate the scalability of the proposed solution.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.38 no.3
• /
• pp.213-220
• /
• 2010

This paper presents the analysis results of trajectory, performance and attitude control based on the first flight data of the KSLV-I. The KSLV-I had a fairing separation problem and failed to inject spacecraft into the orbit. In this paper, the trajectory, flight performance, and attitude control was analyzed considering the influence of unseparated fairing. Moreover, the flight results and performance of the inertial navigation and guidance system were presented. As a results of post-flight analysis, any other problem besides the fairing separation problem was not happened and onboard equipment functioned normally.

• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.11
• /
• pp.1051-1058
• /
• 2012

The unified lateral control algorithm for automatic valet parking for various types of vehicles is presented and its feasibility is shown experimentally via field tests for the given parking scenario. First, a trajectory generation algorithm for forward driving and backward multi-step parking maneuvers is developed. Then, with consideration of different types of vehicles and operating conditions, a kinematic vehicle model is used and validated using field test data. Using the nonlinear vehicle model, the lateral controller is designed based on dynamic surface control. Finally the proposed lateral control law is validated via hardware-in-the-loop simulations for different types of vehicles and experimentally using a test vehicle through field tests.

• Proceedings of the KSR Conference
• /
• 2005.11a
• /
• pp.119-124
• /
• 2005

Decision of operation performance mode to minimize the energy consumption of urban rail vehicle. This paper analyses how much acceleration and deceleration of urban rail vehicle should be applied andhow to choose an operation mode to minimize energy consumption when train runs between station within the fixed operation time. The decided operation pattern satisfying the minimum energy consumption becomes a target trajectory and a basis for the controller design criteria. To make this goal it grasps the characteristics of urban rail vehicle, realize operation energy model of urban rail vehicle and verify the accuracy of embodied model the Matlab simulation with the same operation result of real route. It searches for operation pattern to minimize operation energy by changing the acceleration and deceleration on the imaginative route and proposes operation pattern minimizing energy consumption by applying real operation data between Dolgogee-Sukgye section of Seoul Metropolitan Subway Line 6.