Search Results

Search found 619 results on 25 pages for 'edges'.

Page 14/25 | < Previous Page | 10 11 12 13 14 15 16 17 18 19 20 21  | Next Page >

• Alternative to Daemontools (djbtools) to supervise unix processes?

  - by Stefan Lasiewski

  I've used Daemontools to provide a simple and reliable way to supervise Unix services on my servers. It works well, but it requires a different way of thinking (The DJB Way) and some common complaints are: TAI64N based timestamps Doesn't store scripts under /etc/init.d (or (/usr/local)/etc/rc.d) Doesn't always work with scripts like apachectl. Some scripts need to be rewritten. I remember that some similar "supervisor/watchdog" daemons were in the works about two years ago, but some were still a little rough around the edges. If you have switched from Daemontools to something else, what did you choose and did it work well for you? Does RedHat or Ubuntu come with any process supervisor utilities by default?

  Read the article

• How can I calculate a vertex normal for a hard edge?

  - by K.G.

  Here is a picture of a lovely polygon: Circled is a vertex, and numbered are its adjacent faces. I have calculated the normals of those faces as such (not yet normalized, 0-indexed): Vertex 1 normal 0: 0.000000 0.000000 -0.250000 Vertex 1 normal 1: 0.000000 0.000000 -0.250000 Vertex 1 normal 2: -0.250000 0.000000 0.000000 Vertex 1 normal 3: -0.250000 0.000000 0.000000 Vertex 1 normal 4: 0.250000 0.000000 0.000000 What I'm wondering is, how can I determine, taken as given that I want this vertex to represent a hard edge, whether its normal should be the normal of 1/2 or 3/4? My plan after I glanced at the sketch I used to put this together was "Ha! I'll just use whichever two faces have the same normal!" and now I see that there are two sets of two faces for which this is true. Is there a rule I can apply based on the face winding, angle of the adjacent edges, moon phase, coin flip, to consistently choose a normal direction for this box? For the record, all of the other polygons I plan to use will have their normals dictated in Maya, but after encountering this problem, it made me really curious.

  Read the article

• 2D Rectangle Collision Response with Multiple Rectangles

  - by Justin Skiles

  Similar to: Collision rectangle response I have a level made up of tiles where the edges of the level are made up of collidable rectangles. The player's collision box is represented by a rectangle as well. The player can move in 8 directions. The player's velocity is equal in X and Y directions and constant. Each update, I am checking the player's collision against all tiles that are a certain distance away. When the player collides with a rectangle, I am finding the intersection depth and resolving along the most shallow axis followed by the other axis. This resolution happens for both axes simultaneously. See below for two examples of situations where I am having trouble. Moving up-left against the left wall In the scenario below, the player is colliding with two tiles. The tile intersection depth is equal on both axes for the top tile and more shallow in the X axis for the middle tile. Because the player is moving up the wall, the player should slide in an upward direction along the wall. This works properly as long as the rectangle with the more shallow depth is evaluated first. If the equal intersection depth rectangle is evaluated first, there is a chance the player becomes stuck. Moving up-left against the top wall Here is an identical scenario with the exception that the collision is with the top wall. The same problem occurs at the corners when intersection depth is equal for both axes. I guess my overall question is: How can I ensure that collision response occurs on tiles that have non-equal intersection depth before tiles that have equal intersection depth in order to get around the weirdness that occurs at these corners. Sean's answer in the linked question was good, but his solution required having different velocity components in a certain direction. My situation has equal velocities, so there's no good way to tell which direction to resolve at corners. I hope I have made my explanation clear.

  Read the article

• Check parameters annotated with @Nonnull for null?

  - by David Harkness

  We've begun using FindBugs with and annotating our parameters with @Nonnull appropriately, and it works great to point out bugs early in the cycle. So far we have continued checking these arguments for null using Guava's checkNotNull, but I would prefer to check for null only at the edges--places where the value can come in without having been checked for null, e.g., a SOAP request. // service layer accessible from outside public Person createPerson(@CheckForNull String name) { return new Person(Preconditions.checkNotNull(name)); } ... // internal constructor accessed only by the service layer public Person(@Nonnull String name) { this.name = Preconditions.checkNotNull(name); // remove this check? } I understand that @Nonnull does not block null values itself. However, given that FindBugs will point out anywhere a value is transferred from an unmarked field to one marked @Nonnull, can't we depend on it to catch these cases (which it does) without having to check these values for null everywhere they get passed around in the system? Am I naive to want to trust the tool and avoid these verbose checks? Bottom line: While it seems safe to remove the second null check below, is it bad practice? This question is perhaps too similar to Should one check for null if he does not expect null, but I'm asking specifically in relation to the @Nonnull annotation.

  Read the article

• How do I draw anti-aliased holes in a bitmap

  - by gyozo kudor

  I have an artillery game (hobby-learning project) and when the projectile hits it leaves a hole in the ground. I want this hole to have antialiased edges. I'm using System.Drawing for this. I've tried with clipping paths, and drawing with a transparent color using gfx.CompositingMode = CompositingMode.SourceCopy, but it gives me the same result. If I draw a circle with a solid color it works fine, but I need a hole, a circle with 0 alpha values. I have enabled these but they work only with solid colors: gfx.CompositingQuality = CompositingQuality.HighQuality; gfx.InterpolationMode = InterpolationMode.HighQualityBicubic; gfx.SmoothingMode = SmoothingMode.AntiAlias; In the two pictures consider black as being transparent. This is what I have (zoomed in): And what I need is something like this (made with photoshop): This will be just a visual effect, in code for collision detection I still treat everything with alpha 128 as solid. Edit: I'm usink OpenTK for this game. But for this question I think it doesn't really matter probably it is gdi+ related.

  Read the article

• Convert color photos of documents to good black-and-white images?

  - by Norman Ramsey

  Since I don't have a copier or scanner, I'm using an 8 megapixel camera to copy documents. This works pretty well except they need a lot of processing afterward. I'd like to get from a photo to a bitmap, but using djpeg -grayscale -pnm photo.jpg | pgmtopbm -threshold -value XXX does not work so well, for two reasons: It's hard to guess what XXX should be, and XXX is different for different photos. Illumination varies, and sometimes a single threshold isn't what's right for the image. How can I do better? The ideal solution will be fully automatic command-line program that I can run on Linux. (I have already written a program to remove dark pixels from the edges of images.)

  Read the article

• How to build a "traffic AI"?

  - by Lunikon

  A project I am working on right now features a lot of "traffic" in the sense of cars moving along roads, aircraft moving aroun an apron etc. As of now the available paths are precalculated, so nodes are generated automatically for crossings which themselves are interconnected by edges. When a character/agent spawns into the world it starts at some node and finds a path to a target node by means of a simply A* algorithm. The agent follows the path and ultimately reaches its destination. No problem so far. Now I need to enable the agents to avoid collisions and to handle complex traffic situations. Since I'm new to the field of AI I looked up several papers/articles on steering behavior but found them to be too low-level. My problem consists less of the actual collision avoidance (which is rather simple in this case because the agents follow strictly defined paths) but of situations like one agent leaving a dead-end while another one wants to enter exactly the same one. Or two agents meeting at a bottleneck which only allows one agent to pass at a time but both need to pass it (according to the optimal route found before) and they need to find a way to let the other one pass first. So basically the main aspect of the problem would be predicting traffic movement to avoid dead-locks. Difficult to describe, but I guess you get what I mean. Do you have any recommendations for me on where to start looking? Any papers, sample projects or similar things that could get me started? I appreciate your help!

  Read the article

• XNA 4.0 2D sidescroller variable terrain heightmap for walking/collision

  - by JiminyCricket

  I've been fooling around with moving on sloped tiles in XNA and it is semi-working but not completely satisfactory. I also have been thinking that having sets of predetermined slopes might not give me terrain that looks "organic" enough. There is also the problem of having to construct several different types of tile for each slope when they're chained together (only 45 degree tiles will chain perfectly as I understand it). I had thought of somehow scanning for connected chains of sloped tiles and treating it as a new large triangle, as I was having trouble with glitching at the edges where sloped tiles connect. But, this leads back to the problem of limiting the curvature of the terrain. So...what I'd like to do now is create a simple image or texture of the terrain of a level (or section of the level) and generate a simple heightmap (of the Y's for each X) for the terrain. The player's Y position would then just be updated based on their X position. Is there a simple way of doing this (or a better way of solving this problem)? The main problem I can see with this method is the case where there are areas above the ground that can be walked on. Maybe there is a way to just map all walkable ground areas? I've been looking at this helpful bit of code: http://thirdpartyninjas.com/blog/2010/07/28/sloped-platform-collision/ but need a way to generate the actual points/vectors.

  Read the article

• 3D zooming technique to maintain the relative position of an object on screen

  - by stark

  Is it possible to zoom to a certain point on screen by modifying the field of view and rotating the view of the camera as to keep that point/object in the same place on screen while zooming ? Changing the camera position is not allowed. I projected the 3D pos of the object on screen and remembered it. Then on each frame I calculate the direction to it in camera space and then I construct a rotation matrix to align this direction to Z axis (in cam space). After this, I calculate the direction from the camera to the object in world space and transform this vector with the matrix I obtained earlier and then use this final vector as the camera's new direction. And it's actually "kinda working", the problem is that it is more/less off than the camera's rotation before starting to zoom depending on the area you are trying to zoom in (larger error on edges/corners). It looks acceptable, but I'm not settling for only this. Any suggestions/resources for doing this technique perfectly? If some of you want to explain the math in detail, be my guest, I can understand these things well.

  Read the article

• Anti Aliasing dead

  - by Gazoza

  I have a problem with anti-aliasing. No matter the software settings, it seems to be gone. I tried driver updating and reinstalling, cleaning of the hardware, different monitor with different cablies, OS reinstalling and changing, none of which helped. I decided to change the graphic card, but that didn't work either. Moreover, I have a distinct impression that the jagged edges are worsening as time goes by. I think this is a hardware-related issue, but I don't know exactly what to do. If anyone here had similar troubles, and/or has an idea how to cope with them, I'd be very grateful. My current card is a nVidia GT610.

  Read the article

• Tweaks to make Cleartype better at high resolutions?

  - by ULTRA_POROV

  Cleartype is great when displaying small text (say 10-16px). However when you display something above 20px it starts looking like mud. Just compare it to Photoshop. Photoshop rendering at small size is not very impressive, too blurry. But if you compare it at 20px, Photoshop wins all the time. Cleartype looks jaggy around the edges, almost like there is no Cleartype at all. Can this be fixed, or is it just the way Cleartype is?

  Read the article

• Image manipulation filter needed: unobtrusively hide an area by blurring or smudging?

  - by index

  I would like to hide an empty area of a panorama stitched with hugin (using the GIMP). Hide in the sense of blending it in unobtrusively. I.e. fill the area with the average color of the surroundings and blur it. Or manually smudge the surroundings into the empty area. Is there a filter/plug-in that automatically smudges/blurs the edges into the area? Not looking for seam carving. Thanks.

  Read the article

• What is causing these visual artifacts on my OpenGL sprites?

  - by Amplify91

  What could be the cause of the defects in my characters sprite? I am using OpenGL ES 2.0. I draw my sprites in a sprite batch that uses UV coordinates from one large texture atlas. If you look around the character' edges, you'll see two noticeable problems: The invisible alpha background is not invisible, but shows a strange static-like background. There are unwanted streaks where the character nears the edge of the frame (but only in some frames of the animation, this happened to be one of them). Any idea what could be causing these? I will provide related code if asked for, but I'll try to avoid just dumping the entire project and expecting someone to look through it all. EDIT: Here's a bit of code: This is how I generate my UV coordinates: private float[] createFrameUV(int frameWidth, int frameHeight, int x, int y){ float[] uv = new float[4]; if(numberOfFrames>1){ float width = (float)frameWidth / (float)mBitmap.getWidth(); float height = (float)frameHeight / (float)mBitmap.getHeight(); float u = (float)x / (float)mBitmap.getWidth(); float v = (float)y / (float)mBitmap.getHeight(); uv[0] = u; uv[1] = v; uv[2] = u + width; uv[3] = v + height; }else{ uv[0] = 0f; uv[1] = 0f; uv[2] = 1f; uv[3] = 1f; } return uv; } These are some OpenGL settings: GLES20.glTexParameterf(GLES20.GL_TEXTURE_2D, GLES20.GL_TEXTURE_MIN_FILTER, GLES20.GL_LINEAR); GLES20.glTexParameterf(GLES20.GL_TEXTURE_2D, GLES20.GL_TEXTURE_MAG_FILTER, GLES20.GL_LINEAR); GLES20.glTexParameterf(GLES20.GL_TEXTURE_2D, GLES20.GL_TEXTURE_WRAP_S, GLES20.GL_CLAMP_TO_EDGE); GLES20.glTexParameterf(GLES20.GL_TEXTURE_2D, GLES20.GL_TEXTURE_WRAP_T, GLES20.GL_CLAMP_TO_EDGE);

  Read the article

• Limiting the speed of the mouse cursor

  - by idlewire

  I am working on a simple game where you can drag objects around with the mouse cursor. As I drag the object around quickly, I notice there is some juddering, which seems to be due to the fact that I can move the mouse cursor faster than the game's update/draw. So, although I maintain the offset from where the player initially clicked on the object, the mouse's relative position to the object shifts around slightly before settling as I move the object very quickly. The only way I have found to get smooth, exact 1:1 movement is if I turn both IsFixedTimeStep and SynchronizeWithVerticalRetrace to false. However, I'd rather not have to do that. I have also tried making a custom mouse cursor, hiding the real mouse, taking the real mouse delta and clamping it to a maximum speed. Here is the problem: In windowed mode, the "real" mouse cursor moves off the window while the custom mouse cursor (since it's movement is being scaled) is still somewhere inside the game window. This becomes bizarre and is obviously not desired, as clicking at this point means clicking on things outside the game window. Is there any way to accomplish this in windowed mode? In fullscreen mode, the "real" mouse cursor is bounded to the edges of the screen. So I get to a point where there is no more mouse delta, yet my custom cursor is still somewhere in the middle of the screen and hence can't move further in that direction. If I wanted to clamp it to the edge of the screen when the real cursor is at the edge, then I would get an abrupt jump to the edge of the screen, which isn't desired either Any help would be appreciated. I'd like to be able to limit the speed of the mouse, but also would appreciate help with the first issue (the non-smooth relative offset between mouse cursor movement and object movement).

  Read the article

• Graph data structures and journal format for mini-IDE

  - by matec

  Background: I am writing a small/partial IDE. Code is internally converted/parsed into a graph data structure (for fast navigation, syntax-check etc). Functionality to undo/redo (also between sessions) and restoring from crash is implemented by writing to and reading from journal. The journal records modifications to the graph (not to the source). Question: I am hoping for advice on a decision on data-structures and journal format. For the graph I see two possible versions: g-a Graph edges are implemented in the way that one node stores references to other nodes via memory address g-b Every node has an ID. There is an ID-to-memory-address map. Graph uses IDs (instead of addresses) to connect nodes. Moving along an edge from one node to another each time requires lookup in ID-to-address map. And also for the journal: j-a There is a current node (like current working directory in a shell + file-system setting). The journal contains entries like "create new node and connect to current", "connect first child of current node" (relative IDs) j-b Journal uses absolute IDs, e.g. "delete edge 7 - 5", "delete node 5" I could e.g. combine g-a with j-a or combine g-b with j-b. In principle also g-b and j-a should be possible. [My first/original attempt was g-a and a version of j-b that uses addresses, but that turned out to cause severe restrictions: nodes cannot change their addresses (or journal would have to keep track of it), and using journal between two sessions is a mess (or even impossible)] I wonder if variant a or variant b or a combination would be a good idea, what advantages and disadvantages they would have and especially if some variant might be causing troubles later.

  Read the article

• Interesting 3d zooming technique

  - by stark

  Is it possible to zoom to a certain point on screen by modifying the field of view and rotating the camera as to keep that point/object in the same place on screen while zooming ? Changing the camera position is not allowed.. I projected the 3d pos of the object on screen and remembered it. Then on each frame I calculate the direction to it in camera space and then I construct a rotation matrix to align this direction to Z axis (in cam space). After this, I calculate the direction from the camera to the object in world space and transform this vector with the matrix I obtained earlier and then use this final vector as the camera's new direction. And it's actually "kinda working", the problem is that it is more/less off than the camera's rotation before starting to zoom depending on the area you are trying to zoom in (larger error on edges/corners). It looks acceptable, but I'm not settling for only this. Any suggestions/resources for doing this technique perfectly ? If some of you want to explain the math in detail, be my guests, I can understand these things well. Thanks. Edit: I'll check often for responses, I'm really curious about this :D

  Read the article

• Who can change the View in MVC?

  - by Luke

  I'm working on a thick client graph displaying and manipulation application. I'm trying to apply the MVC pattern to our 3D visualization component. Here is what I have for the Model, View, and Controller: Model - The graph and it's metadata. This includes vertices, edges, and the attributes of each. It does not contain position information, icons, colors, or anything display related. View - This would commonly be called a scene graph. It includes the 3D display information, texture information, color information, and anything else that is related specifically to the visualization of the model. Controller - The controller takes the view and displays it in a Window using OpenGL (but it could potentially be any 3D graphics package). The application has various "layouts" that change the position of the vertices in the display. For instance, one layout may arrange the vertices in a circle. Is it common for these layouts to access and change the view directly? Should they go through the Controller to access the View? If they go through the Controller, should they just ask for direct access to the View or should each change go through the controller? I realize this is a bit different from the standard MVC example where there a finite number of Views. In this case, the View can change in an infinite number of ways. Perhaps I'm shattering some basic principle of MVC here. Thanks in advance!

  Read the article

• Creating a 2D perspective in 3D game

  - by Accatyyc

  I'm new to XNA and 3D game development in general. I'm creating a puzzle game kind of similar to tetris, built with blocks. I decided to build the game in 3D since I can do some cool animations and transitions when using 3D blocks with physics etc. However, I really do want the game to look "2D". My blocks are made up of 3D models, but I don't want that to be visible when they're not animating. I have followed some XNA tutorials and set up my scene like this: this.view = Matrix.CreateLookAt(cameraPosition, Vector3.Zero, Vector3.Up); this.aspectRatio = graphics.GraphicsDevice.Viewport.AspectRatio; this.projection = Matrix.CreatePerspectiveFieldOfView( MathHelper.ToRadians(45.0f), aspectRatio, 1.0f, 10000.0f); ... and it gives me a very 3D-ish look. For example, the blocks in the center of the screen looks exactly how I want them, but closer to the edges of the screen I can see the rotation and sides of them. My guess is that I'm not after a perspective field of view, but any help on which field of view/settings to use to get a "flat" look when the blocks aren't rotated would be great!

  Read the article

• Why isn't my other two constant buffers being updated to the shader?

  - by Paul Ske

  I posted previously before about my two dynamic buffers not being dynamically updating the constant shader. The tessellation buffer isn't working because I have to manually update the tessellation factor inside the hull shader. I believe the camera position isn't updating either because when I perform distance adaptation the far edges are more tessellated then the what's truly in front of the camera. I have all the buffers set to dynamic. Inside the render loop I have them set as: ID3D11Buffer *multiBuffers[3]; devcon->VSSetConstantBuffers(0,3,multiBuffers); ... devcon->DSSetConstantBuffers(0,3,multiBuffers); I only got that from a directX Sample. Inside the shader file I have the three cbuffer structs. cbuffer ConstantBuffer { float4x4 WorldMatrix; float4x4 viewMatrix; float4x4 projectionMatrix; float4x4 modelWorldMatrix; // the rotation matrix float3 lightvec; // the light's vector float4 lightcol; // the light's color float4 ambientcol; // the ambient light's color bool isSelected; } cbuffer cameraBuffer { float3 cameraDirection; float padding; } cbuffer TessellationBuffer { float tessellationAmount; float3 padding2; } Am I missing something or would anyone know why wouldn't my buffers update to the shader file?

  Read the article

• a flexible data structure for geometries

  - by AkiRoss

  What data structure would you use to represent meshes that are to be altered (e.g. adding or removing new faces, vertices and edges), and that have to be "studied" in different ways (e.g. finding all the triangles intersecting a certain ray, or finding all the triangles "visible" from a given point in the space)? I need to consider multiple aspects of the mesh: their geometry, their topology and spatial information. The meshes are rather big, say 500k triangles, so I am going to use the GPU when computations are heavy. I tried using arrays with vertices and arrays with indices, but I do not love adding and removing vertices from them. Also, using arrays totally ignore spatial and topological information, which I may need studying the mesh. So, I thought about using custom double-linked list data structures, but I believe doing so will require me to copy the data to array buffers before going on the GPU. I also thought about using BST, but not sure it fits. Any help is appreciated. If I have been too fuzzy and you require other information feel free to ask.

  Read the article

• How can I determine the first visible tile in an isometric perspective?

  - by alekop

  I am trying to render the visible portion of a diamond-shaped isometric map. The "world" coordinate system is a 2D Cartesian system, with the coordinates increasing diagonally (in terms of the view coordinate system) along the axes. The "view" coordinates are simply mouse offsets relative to the upper left corner of the view. My rendering algorithm works by drawing diagonal spans, starting from the upper right corner of the view and moving diagonally to the right and down, advancing to the next row when it reaches the right view edge. When the rendering loop reaches the lower left corner, it stops. There are functions to convert a point from view coordinates to world coordinates and then to map coordinates. Everything works when rendering from tile 0,0, but as the view scrolls around the rendering needs to start from a different tile. I can't figure out how to determine which tile is closest to the upper right corner. At the moment I am simply converting the coordinates of the upper right corner to map coordinates. This works as long as the view origin (upper right corner) is inside the world, but when approaching the edges of the map the starting tile coordinate obviously become invalid. I guess this boils down to asking "how can I find the intersection between the world X axis and the view X axis?"

  Read the article

• How to read reduce/shift conflicts in LR(1) DFA?

  - by greenoldman

  I am reading an explanation (awesome "Parsing Techniques" by D.Grune and C.J.H.Jacobs; p.293 in the 2nd edition) and I moved forward from my last question: How to get lookahead symbol when constructing LR(1) NFA for parser? Now I have such "problem" (maybe not a problem, but rather need of confirmation from some more knowledgeable people). The authors present state in LR(0) which has reduce/shift conflict. Then they build DFA for LR(1) for the same grammar. And now they say it does not have a conflict (lookaheads at the end): S -> E . eof E -> E . - T eof E -> E . - T - and there is an edge from this state labeled - but no labeled eof. Authors says, that on eof there will be reduce, on - there will be shift. However eof is for shift as well (as lookahead). So my personal understanding of LR(1) DFA is this -- you can drop lookaheads for shifts, because they serve no purpose now -- shifts rely on input, not on lookaheads -- and after that, remove duplicates. S -> E . eof E -> E . - T So the lookahead for reduce serves as input really, because at this stage (all required input is read) it is really incoming symbol right now. For shifts, the input symbols are on the edges. So my question is this -- am I actually right about dropping lookaheads for shifts (after fully constructing DFA)?

  Read the article

• Recommended display/background brightness ratio and UI color schemes [duplicate]

  - by user1306322

  This question already has an answer here: Colour scheme for editor - guidelines or medical reccomendations 3 answers I'm a professional programmer, which means I spend a lot of time staring at various displays. Recently I've been having some problems with my eyes, so I went to talk to several doctors, which all gave me different recommendations as to how bright the background of the room should be in comparison to the display's brightness. It was very confusing, as some of them even agreed with counter-arguments of others, which made it all even less clear. So I'd like to ask the professional programmers, as people who actually have some experience with that. Some of the doctors said that looking at a monitor is like looking at a book, so the brightness ratios should be approximately the same. Others said that background should be as bright as the display itself, because then there is no brightness difference at the edges, and that's what may cause eye fatigue. From my own experience, I can say that reading a book isn't the same as writing or debugging a program, where you have to pay close attention to each symbol, and in books most words are easily recognizable without focusing too hard on them. Also, books are black on white and I myself use the default (dark text, white bg) color scheme for my IDE, but I've seen some programmers use mid-bright text on very dark background color schemes. So I'd like to ask what are the recommended display/background brightness ratios for programming? I'm not sure this site is the right one for this kind of questions, so if you know a better one, please comment.

  Read the article

• Spacebar on Samsung Series 7 Chronos Not Working

  - by Samg

  The spacebar on my Series 7 Chronos does not register half the time. I'll be typing along and my words end up lookinglikethis. After further investigation, I noticed the spacebar does not response when pressed at the edges. Basically, the areas under the "C" key and the "M" key do not respond to a press It's weird, the bar physically "clicks" down (like it should), but does not register. No matter how hard I press, It won't register in that area. I noticed some complaints online about a 'sticky' or 'loose' spacebar, but they didn't mention it not registering. We took it back to the store we bought it at and they noticed the same problem on the floor model. Does anyone know how to fix this?

  Read the article

• How to build a "traffic AI"?

  - by Lunikon

  A project I am working on right now features a lot of "traffic" in the sense of cars moving along roads, aircraft moving aroun an apron etc. As of now the available paths are precalculated, so nodes are generated automatically for crossings which themselves are interconnected by edges. When a character/agent spawns into the world it starts at some node and finds a path to a target node by means of a simply A* algorithm. The agent follows the path and ultimately reaches its destination. No problem so far. Now I need to enable the agents to avoid collisions and to handle complex traffic situations. Since I'm new to the field of AI I looked up several papers/articles on steering behavior but found them to be too low-level. My problem consists less of the actual collision avoidance (which is rather simple in this case because the agents follow strictly defined paths) but of situations like one agent leaving a dead-end while another one wants to enter exactly the same one. Or two agents meeting at a bottleneck which only allows one agent to pass at a time but both need to pass it (according to the optimal route found before) and they need to find a way to let the other one pass first. So basically the main aspect of the problem would be predicting traffic movement to avoid dead-locks. Difficult to describe, but I guess you get what I mean. Do you have any recommendations for me on where to start looking? Any papers, sample projects or similar things that could get me started? I appreciate your help!

  Read the article

< Previous Page | 10 11 12 13 14 15 16 17 18 19 20 21  | Next Page >