• Journal of Digital Convergence
• /
• v.17 no.10
• /
• pp.233-242
• /
• 2019

Numerous space perception studies have shown that Euclidean 3-D structure cannot be recovered from binocular stereopsis, motion, combination of stereopsis and motion, or even with combined multiple sources of optical information. Humans, however, have no difficulties to perform the task-specific action despite of poor shape perception. We have applied humans skill and capabilities to artificial intelligence and computer vision but those machines are still far behind from humans abilities. Thus, we need to understand how we perceive depth in space and what information we use to perceive 3-D structure accurately to perform. The purpose of this paper was to review space perception literatures to apply humans abilities to artificial intelligence robots more advanced in future.

• ETRI Journal
• /
• v.45 no.5
• /
• pp.836-846
• /
• 2023

Mobile robots are used in modern life; however, object recognition is still insufficient to realize robot navigation in crowded environments. Mobile robots must rapidly and accurately recognize the movements and shapes of pedestrians to navigate safely in pedestrian-rich spaces. This study proposes real-time, accurate, three-dimensional (3D) multi-pedestrian detection and tracking using a 3D light detection and ranging (LiDAR) point cloud in crowded environments. The pedestrian detection quickly segments a sparse 3D point cloud into individual pedestrians using a lightweight convolutional autoencoder and connected-component algorithm. The multi-pedestrian tracking identifies the same pedestrians considering motion and appearance cues in continuing frames. In addition, it estimates pedestrians' dynamic movements with various patterns by adaptively mixing heterogeneous motion models. We evaluate the computational speed and accuracy of each module using the KITTI dataset. We demonstrate that our integrated system, which rapidly and accurately recognizes pedestrian movement and appearance using a sparse 3D LiDAR, is applicable for robot navigation in crowded spaces.

• The Journal of the Korea Contents Association
• /
• v.21 no.1
• /
• pp.541-551
• /
• 2021

Among the numerous duties of a librarian in a library, the work of arranging books is a job that the librarian has to do one by one. Thus, the cost of labor and time is large. In order to solve this problem, the interest in book-arranging robots based on artificial intelligence has recently increased. In this paper, we propose the K-ACO algorithm, which is the shortest path algorithm for multi-stops that can be applied to the library book arrangement robots. The proposed K-ACO algorithm assumes multiple robots rather than one robot. In addition, the K-ACO improves the ANT algorithm to create K clusters and provides the shortest path for each cluster. In this paper, the performance analysis of the proposed algorithm was carried out from the perspective of book arrangement time. The proposed algorithm, the K-ACO algorithm, was applied to a university library and compared with the current book arrangement algorithm. Through the simulation, we found that the proposed algorithm can allocate fairly, without biasing the work of arranging books, and ultimately significantly reduce the time to complete the entire work. Through the results of this study, we expect to improve quality services in the library by reducing the labor and time costs required for arranging books.

• International Journal of Precision Engineering and Manufacturing
• /
• v.7 no.1
• /
• pp.62-66
• /
• 2006

The learning control develops controllers that learn to improve their performance at executing a given task, based on experience performing this specific task. In a previous work, the authors presented an iterative precision of linear decentralized learning control based on p-integrated learning method for the vertical dynamic multiple systems. This paper develops an indirect decentralized learning control based on adaptive control method. The original motivation of the learning control field was learning in robots doing repetitive tasks such as an assembly line works. This paper starts with decentralized discrete time systems, and progresses to the robot application, modeling the robot as a time varying linear system in the neighborhood of the nominal trajectory, and using the usual robot controllers that are decentralized, treating each link as if it is independent of any coupling with other links. Some techniques will show up in the numerical simulation for vertical dynamic robot. The methods of learning system are shown for the iterative precision of each link.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.8
• /
• pp.738-744
• /
• 2013

According to the cognitive science research, the interaction intent of humans can be estimated through an analysis of the representing behaviors. This paper proposes a novel methodology for reliable intention analysis of humans by applying this approach. To identify the intention, 8 behavioral features are extracted from the 4 characteristics in human-human interaction and we outline a set of core components for nonverbal behavior of humans. These nonverbal behaviors are associated with various recognition modules including multimodal sensors which have each modality with localizing sound source of the speaker in the audition part, recognizing frontal face and facial expression in the vision part, and estimating human trajectories, body pose and leaning, and hand gesture in the spatial part. As a post-processing step, temporal confidential reasoning is utilized to improve the recognition performance and integrated human model is utilized to quantitatively classify the intention from multi-dimensional cues by applying the weight factor. Thus, interactive robots can make informed engagement decision to effectively interact with multiple persons. Experimental results show that the proposed scheme works successfully between human users and a robot in human-robot interaction.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.4
• /
• pp.92-98
• /
• 2005

The learning control develops controllers that learn to improve their performance at executing a given task, based on experience performing this specific task. In a previous work, the authors presented an iterative precision of linear decentralized learning control based on p-integrated learning method for the vertical dynamic multiple systems. This paper develops an indirect decentralized teaming control based on adaptive control method. The original motivation of the teaming control field was loaming in robots doing repetitive tasks such as on an assembly line. This paper starts with decentralized discrete time systems, and progresses to the robot application, modeling the robot as a time varying linear system in the neighborhood of the nominal trajectory, and using the usual robot controllers that are decentralized, treating each link as if it is independent of any coupling with other links. Some techniques will show up in the numerical simulation for vertical dynamic robot. The methods of learning system are shown up for the iterative precision of each link.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.1
• /
• pp.42-47
• /
• 2015

This paper presents a rank-based formation for multiple agents based on potential functions, where the proposed method uses the relative position of two neighboring agents. The conventional formation scheme of multiple systems requires communication between agents and a central computer to get the positions of all multiple agents. In the study, differently from previous studies, the formation scheme uses the relative position of two neighboring agents in a local coordinate system. In addition, it introduces a singular agent association that considers only the relative position between an agent and its neighboring agents, instead of multiple associations among all information about all agents. Furthermore, the proposed framework explores the benefits of different formation types. Extensive simulation results show that the proposed approach verifies the viability and effectiveness of the proposed formation.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.17 no.5
• /
• pp.643-653
• /
• 2014

High-capacity video, control and situation awareness data should be transmitted efficiently to control robots properly in the ground unmanned system, which requires the technology maximizing the communication range and the data transmission throughput. This technology is connected to the OFDMA(Orthogonal Frequency Division Multiple Access)-MIMO(Multiple Input Multiple Output) transmission technology under the limited bandwidth and transmission power. In this paper, we design MIMO communication system for ground unmanned systems, and investigate the data reception performance experimentally, comparing with SISO(Single Input Single Output) system. Experiment results show that the data reception performance of MIMO is significantly improved compared to that of SISO, e.g. 4dB gain of sensitivity and 5dB of SNR at the value MPR = 1, for the mobile stations with $2{\times}2$ STBC diversity.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.74-79
• /
• 2005

In this paper, we propose a novel approach to decentralized motion planning and conflict-resolution for multiple mobile agents working in an environment with unexpected moving obstacles. Our proposed motion planner has two characteristics. One is a real-time collision prognosis based on modified collision map. Collision map is a famous centralized motion planner with low computation load, and the collision prognosis hands over these characteristics. And the collision prognosis is based on current robots status, maximum robot speeds, maximum robot accelerations, and path information produced from off-line path planning procedure, so it is applicable to motion planner for multiple agents in a dynamic environment. The other characteristic is that motion controller architecture is based on potential field method, which is capable of integrating robot guidance to the goals with collision avoidance. For the architecture, we define virtual obstacles making delay time for collision avoidance from the real-time collision prognosis. Finally the results obtained from realistic simulation of a multi-robot environment with unknown moving obstacles demonstrate safety and efficiency of the proposed method.

• The Journal of Korea Robotics Society
• /
• v.17 no.4
• /
• pp.407-416
• /
• 2022

Multiple AMRs have been proved to be effective in improving warehouse productivity by eliminating workers' wasteful walking time. Although Multi-agent Path Finding (MAPF)-based solution is an optimal approach for this task, its deployment in practice is challenging mainly due to its imperfect plan-execution capabilities and insufficient computing resources for high-density environments. In this paper, we present a MAPF-based fleet management system architecture that robustly manages multiple robots by re-computing their paths whenever it is necessary. To achieve this, we defined four events that trigger our MAPF solver framework to generate new paths. These paths are then delivered to each AMR through ROS2 message topic. We also optimized a graph structure that effectively captures spatial information of the warehouse. By using this graph structure we can reduce computational burden while keeping its rescheduling functionality. With proposed MAPF-based fleet management system, we can control AMRs without collision or deadlock. We applied our fleet management system to the real logistics warehouse with 10 AMRs and observed that it works without a problem. We also present the usage statistic of adopting AMRs with proposed fleet management system to the warehouse. We show that it is useful over 25% of daily working time.