Abstract:
PURPOSE: An illegal user detecting device and a method thereof are provided to cope with the change of a gamebot or an auto program by detecting an illegal user based on the differences of unique characteristics among the gamebot, the auto program, and a normal user. CONSTITUTION: A user characteristic analysis unit (100) analyzes unique characteristic differences between a normal user and an illegal program. A recognizing method determining unit (102) determines a method for recognizing the illegal program based on the unique characteristic differences. An illegal user detection unit (106) detects the illegal user who uses the illegal program in a game based on the recognizing method. The user characteristic analysis unit analyzes the unique characteristic differences based on the properties of the illegal program and the game log data of a normal user. [Reference numerals] (100) User characteristic analysis unit; (102) Recognizing method determining unit; (104) Verification unit; (106) Illegal user detection unit; (108) Detecting result corresponding unit; (110) Detecting result tracing unit; (112) Database; (AA) Reference data
Abstract:
PURPOSE: A real-time route searching device and a method are provided to deal with emergency situations even while moving to a natural route because an optimized route searching guide is generated and real-time change is reflected. CONSTITUTION: A real-time route searching device comprises an input unit(101), an extracting unit(103), and an analyzing unit(104). Information related to a map and real-time change information is input to the input unit. The extracting unit extract route searching information required for performing a route search from the information related to the map and real-time change information. The analyzing unit generates a route searching guide by analyzing the information related the map and a route searching path reflecting the real-time change information to the route search guide.
Abstract:
PURPOSE: A system and method for providing a virtual reality linking service is provided to enable a user to freely convert a real life and a virtual life by reconstructing and connecting the virtual life, the real life, and the user. CONSTITUTION: A receiver(101) receives one or more information among actual life object property information, actual life object movement information, and establishment information by object. A virtual object manager(111) creates and establishes the virtual space. The target object manager(113) creates and manages one or more objects.
Abstract:
본 발명은 공간 분할을 이용한 게임 캐릭터의 패턴 학습 기술에 관한 것으로, 게임 내 캐릭터가 사용자의 행동 양식을 학습할 수 있도록 하여 사용자와 유사한 행동을 수행하기 위해 게임의 물리적, 논리적, 캐릭터 별 상태 공간을 분할하고 각각을 신경망으로 분리하여 학습을 수행하도록 한다. 그리고 게임 캐릭터를 신경망으로 학습시킬 때, 신경망의 출력을 확률로 계산하여 게임에 적용하는 것을 특징으로 한다. 본 발명에 의하면, 게임의 상태 공간 분할을 통하여 학습 시간과 결과 측면에서 성능 향상을 꾀할 수 있으며, 신경망의 출력노드 값이 크기에 비례하여 출력노드가 선택될 수 있도록 하여 게임에서 캐릭터가 다양한 행동을 수행할 수 있으며, 더 나아가 게임 유저와 비슷한 행동을 할 수 있다. 공간분할, 게임 캐릭터, 패턴 학습
Abstract:
PURPOSE: A virtual client management module supply device and a game test method using the same are provided to efficiently manage the virtual client by dividing commonly applied matters and differently applied matters. CONSTITUTION: A specialized module generator(110) create a specialized module. The specialized module expands the specialized templates by the sorts of game using the intrinsic information of the game. A common module provider(120) supplies a common module including common functions for the game. A module combiner(130) creates the virtual client management module of the game by combining the specialized module and the common module.
Abstract:
PURPOSE: A packet capture analysis device based on a game grammar for a game test and a method thereof are provided to implement the load test of a server and the general quality verification of a game. CONSTITUTION: A packet capturing unit(102) captures the packet of game data between a game client(100) and a game server(150). Based on an analyzed result, a packet analysis unit(106) creates a game grammar through the analysis of the captured packet of the packet capturing unit. A game grammar generator(114) creates analyzed packet generation rules, a protocol, an action, etc. to the game grammar. According to the game grammar, a load generating unit(116) creates the packet. The loading generating unit generates load to the game server.
Abstract:
PURPOSE: An artificial intelligence production method of a game character is provided to create the artificial intelligence of the game character easily by applying a learning algorithm to inter-state figurative meaning after the creation of the finite-state machine of a basic character. CONSTITUTION: In case an artificial intelligence reasoning begins, a recognition unit offers a game state to a finite state machine generator(S203,S205). The finite state machine generator creates the finite-state machine of a basic character(S207). A result generator creates automatically the artificial intelligence of various game characters through the study result of a learning unit(S211). A delivery unit transfers the artificial intelligence to game contents(S213).
Abstract:
A method and a system for rigid body simulation are provided to enable a time-division controller to adjust the accuracy and stability of ration by adjusting a visual interval in an LMS(Largrange Multiplier Simulation) state and an IBS(Impulse Based Simulation) state. A rigid body information initializing unit(100) forms an initial position, pose, mass and inertia tensor of a rigid body and generates joint connection information. A force collecting unit(102) collects an external force added to the rigid body. A collision inspecting unit(104) performs a collection inspection between rigid bodies. An LMS unit(106) calculates a rigid body state after the time, a*Deltat(a is a variable of Deltat, and a>=0), and an IBS unit calculates a rigid body state after the time, b*Deltat(b is a variable of Deltat, and b>=0). A time division control unit(110) divides the simulation time interval into the a*Deltat and the b*Deltat.
Abstract:
A multi-threading framework for supporting the dynamic load balancing and a processing method using the same are provided to offer a dynamic load balancing function by using the number of thread pools, option levels of unit task and run time option levels. According to unit task information transmitted from a specific application, a task scheduler(202) re-defines the processing order. A device enumerator(204) detects a device in which the specific application is performed. The resource used in the application of the detected device is defined. A resource manager(208) manages resources. A plug-in manager(210) manages a plurality of modules which perform various functions related to the specific application in a plug-in form.
Abstract:
A method and an apparatus for extracting a line segment are provided to obtain the line segment through weight based line fitting by using found pixels. A direction image and a slop image for each edge pixel from a canny girth image obtained from an inputted image are calculated(S204). Arbitrary two pixels of the girth pixels are selected to form a straight line. A line segment candidate pixel is selected through gradual pixel extension from a central point of the two pixels forming the straight line. After performing weight based line fitting by calculating weight values of the selected line segment candidate pixels, a line segment is extracted by checking whether the line segment candidate pixels are connected or not.