• Journal of Korea Multimedia Society
• /
• v.24 no.1
• /
• pp.1-12
• /
• 2021

In this paper, we present a new architecture for semantic segmentation. Semantic segmentation aims at a pixel-wise classification which is important to fully understand images. Previous semantic segmentation networks use features of multi-layers in the encoder to predict final results. However, they do not contain various receptive fields in the multi-layers features, which easily lead to inaccurate results for boundaries between different classes and small objects. To solve this problem, we propose a multi-path feature fusion module that allows for features of each layers to contain various receptive fields by use of a set of dilated convolutions with different dilatation rates. Various experiments demonstrate that our method outperforms previous methods in terms of mean intersection over unit (mIoU).

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.7
• /
• pp.1924-1937
• /
• 1996

In this paper, we propose a ESP(Extended Scan Path) architecture for multi- board testing. The conventional architectures for board testing are single scan path and multi-scan path. In the single scan path architecture, the scan path for test data is just one chain. If the scan path is faulty due to short or open, the test data is not valid. In the multi-scan path architecture, there are additional signals in multi-board testing. So conventional architectures are not adopted to multi-board testing. In the case of the ESP architecture, even though scan paths either short or open, it doesn't affect remaining other scan paths. As a result of executing parallel BIST and IEEE 1149.1 boundary scan test by using, he proposed ESP architecture, we observed to the test time is short compared with the single scan path architecture. Because the ESP architecture uses the common bus, there are not additional signals in multi-board testing. By comparing the ESP architecture with conventional one using ISCAS '85 bench mark circuit, we showed that the architecture has improved results.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.5
• /
• pp.416-422
• /
• 2013

This paper suggests a hybrid path planning method of multi-robots, where a path deviation prevention for maintaining a specific formation is implemented by using repulsive function, $A^*$ algorithm and UKF (Unscented Kalman Filter). The repulsive function in potential field method is used to avoid collision among robots and obstacles. $A^*$ algorithm helps the robots to find optimal path. In addition, error estimation based on UKF guarantees small path deviation of each robot during navigation. The simulation results show that the swarm robots with designated formation successfully avoid obstacles and return to the assigned formation effectively.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.15 no.3 s.96
• /
• pp.280-289
• /
• 2005

A single degree of freedom system and transmissibility are key concepts in many problems of vibration isolation. In order to apply this approach, however, several assumptions must be satisfied, which are often not realistic. For examples, in practical systems, vibration transmissions at multi-point with multi-degree of freedom(translational and rotational DOF) take place and mobilities or impedances of receiver structures cannot be ignored any more especially over high frequency range. Therefore, a multi-dimensional treatment is required for accurate estimation of dynamic behavior of the system. In this paper, an approach using vibrational power flow is introduced to deal with analysis of multi-dimensional vibration isolation system in a more practical way and in aspects of vibration isolations and vibration path analysis. Procedures of this approach and some results of research for vibrational power path analysis with rotational terms included are presented. Difficulties in this method are also discussed.

• Journal of KIISE:Computing Practices and Letters
• /
• v.7 no.1
• /
• pp.12-23
• /
• 2001

Extended path expressions are used to denote multiple paths concisely by using '$\ast$' character. They are convenient for expressing OQL queries to retrieve XML data stored in OODBs. In this paper, we propose a multi-path index scheme as a new index scheme to efficiently process queries with extended path expressions. Our proposed index scheme allocates a unique path identifier for every possible single path in an extended path expression and provides functionalities of both a single path indexing and multiple path indexing through the composition of index key and path identifier while using only a index structure. The proposed index scheme provides better performance than single-path index schemes, and is practical since it can be implemented by little modification of leaf records of a B+-tree index.

• Journal of KIISE:Information Networking
• /
• v.35 no.4
• /
• pp.292-300
• /
• 2008

In multi-hop wireless networks like Wireless Mesh Networks (WMN) and Wireless Sensor Networks (WSN), nodes often rely on batteries as their power source. In such cases, energy efficient routing is critical. Many schemes have been proposed to find the most energy efficient path, but most of them do not achieve optimality on network lifetime. Once found, the energy efficient path is constantly used such that the energy of the nodes on the path is depleted quickly. As an alternative, the approaches that dynamically change the path at run time have also been proposed. These approaches, however, involve high overhead of establishing multiple paths. In this paper, we first find an optimal multi-path routing using LP. Then we apply an approximation algorithm to derive a near-optimal solution for single-path routing. We compare the performance of the proposed scheme with several other existing algorithms through simulation.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.1
• /
• pp.11-19
• /
• 2013

For an autonomous multi-mobile robot system, path planning and collision avoidance are important functions used to perform a given task collaboratively and cooperatively. This study considers these important and challenging problems. The proposed approach is based on a potential field method and fuzzy logic system. First, a global path planner selects the paths of the robots that minimize the cost function from each robot to its own target using a potential field. Then, a local path planner modifies the path and orientation from the global planner to avoid collisions with static and dynamic obstacles using a fuzzy logic system. In this paper, each robot independently selects its destination and considers other robots as dynamic obstacles, and there is no need to predict the motion of obstacles. This process continues until the corresponding target of each robot is found. To test this method, an autonomous multi-mobile robot simulator (AMMRS) is developed, and both simulation-based and experimental results are given. The results show that the path planning and collision avoidance strategies are effective and useful for multi-mobile robot systems.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.05a
• /
• pp.702-704
• /
• 2010

Wireless LAN(Local Area Network) environment to support QoS in IEEE802.11 Working Group has standardized on the IEEE 802.11e. IEEE 802.11e EDCA channel access of the high priority traffic as by raising the probability of a chance to support QoS. This is a high priority traffic than low priority traffic, channel access is not necessarily the first. Therefore, as all traffic channels, if you have the same routing path, lower priority traffic on the performance of high priority traffic, sustainable impact. As a result, EDCA, such as voice or video to guarantee QoS for multimedia applications, follow the crowd. In this paper, to solve these problems by using Multi-Interface Multi-Path Routing Algorithm is proposed. Interface for each use in IEEE 802.11e EDCA Access Category much differently by each of the Routing Path for high priority traffic and guarantee QoS.

• The KIPS Transactions:PartC
• /
• v.15C no.4
• /
• pp.289-296
• /
• 2008

In wireless mesh network, studies on routing protocols have been actively carried out recently, and hop count is used as a major routing metric in destination-sequenced distance-vector (DSDV) routing protocol, which is a representative proactive routing protocol. Although hop-by-hop multi-path (HMP) DSDV and enhanced HMP (EHMP) DSDV routing protocols perform routing by considering both hop count and residual bandwidth within one hop distance nodes, it has a shortcoming that routing is carried out via non-optimal path from the aspect of end-to-end routing. In order to overcome the shortcoming, a cost-aware multi-path (CAMP) DSDV routing protocol is proposed in this paper, which considers hop count and end-to-end minimum residual bandwidth. Simulation results based on NS-2 show that the proposed routing protocol performs better than DSDV, HMP DSDV, and EHMP DSDV protocols from the aspect of throughput and packet delivery ratio, by appropriately using hop count and end-to-end minimum residual bandwidth information and has the same number of management messages with HMP DSDV and EHMP DSDV protocols.

• Ocean and Polar Research
• /
• v.36 no.4
• /
• pp.465-473
• /
• 2014

This paper proposes a formation control method by which multi-mining robots maintain a specified formation and follow a path. To secure the path tracking performance, a pure-pursuit algorithm is considered for each individual robot, and to minimize the deviation from the reference path, speed reduction in the steering motion is added. For the formation, in which two robots are parallel in a lateral direction, the robots track the specified path at a constant distance. In this way, the Leader-Follower method is adopted and the following robot knows the position and heading angle of the leader robot. Through the experimental test using two ground vehicle models, the performance is verified.