presentation5

Post on 16-Apr-2017

103 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

APAPTIVE APAPTIVE MOTION MOTION

PLANNING IN PLANNING IN CROWD CROWD

BEHAVIOURBEHAVIOURAkshay Anand & Himanshu Bhatia

CSE-2, 8th Sem

GOALGOAL• Simulate flow of traffic as crowd behaviors on road

and implement characteristics such as:

Define path for movement Apply Queuing Apply collision avoidance Path finding technique

• We plan to show different classes of behavior in autonomous entities, and investigate the difference in their paths, speeds.

Different Types of Vehicles

• In a real life situation, all driven vehicles are not equal in their behavior.

1. Some may follow the rules.2. Some may be moderate.3. For some, speed is the ultimate thrill.

Traffic Simulation vs. Traffic Simulation vs. Crowd SimulationCrowd Simulation

• Already published work simulates traffic flow, but in an environment where the traffic rules such as lanes are observed.

• Incorporate fundamentals of crowd behavior in our traffic simulation, so that different classes of agents behave differently.

Crowd SimulationCrowd Simulation• Process of simulating the movement of a large

number of objects or characters.

• Observed human behavior interaction is taken into account, to replicate the collective behavior.

Traffic as CrowdTraffic as Crowd• In India traffic movement is haphazard, and so is

natural crowd movement.

• Consider each vehicle as a person, and each vehicle carries a degree of o Aggressivenesso Non-compliance to rules

Hardware & Software Hardware & Software requiredrequired

• Models and simulations may require high graphics capabilities (3D photorealism) and be capable of processing high end mathematical models which can be very CPU intensive. Machines like this will require:

• a fast CPU (at least 2.0 Ghz )• large amounts of RAM( at least 2 Gb)• A good graphics card( DirectX 9 compatible cards)• large storage capacity (i.e. large hard drive)• may require specialised input output devices• will require specialised software (XNA /OpenGL)• Languague: C#/c++

ModulesModuleswe have incorporated in the we have incorporated in the

projectproject • Define terrain and vehicle /agent• Define Path finding technique.• Define path for movement• Apply Queuing• Apply collision avoidance

WE HAVE USED C# AS THE LANGUAGUE AND XNA FRAMEWORK TO DEVELOP THE SIMULATION .

TerrainTerrainThis is a sample road texture what we have used as the background. The cars will be moving inside the enclosed area. 1000x 300 PIXELS

Agent/vehiclesAgent/vehiclesA 2D model what we have implemented in the project.

This is a 3D model made using effect file in direct X.

DEFINING THE DEFINING THE CHARACTERISTICS OF CHARACTERISTICS OF

THE VEHICLETHE VEHICLEVariables defined for

implementingphysics:

1. SPEED2. ACCELERATION,GEAR,3. STEERANGLE,4. TOPSPEEDS, 5. MAXREVERSESPEED

Path finding-using A* Path finding-using A* AlgorithmAlgorithm

• Path planning consists of finding an optimalpath (generally the shortest one) between astarting point and a destination point in a virtual environment, avoiding obstacles.

Adapting motion according Adapting motion according to behaviorto behavior

• The A-Star algorithm in our program works according to heuristics defined by us.

• H= p1*(speed)+ p2*(next lane distance)

• For constants p1 and p2:• Class 1: p2>>>>p1• Class 2 : p1 = p2• Class 3: p1>>>>p2

• Vehicle will define its path according to least measure of H for next pixel.

Path calculationPath calculation

Queuing depends onQueuing depends on• Politeness and panic• Waiting rule applied (similar

with stopping rule) • Influence of one agent on other.

Organized Behavior : Organized Behavior : QueuingQueuing

Collision avoidanceCollision avoidance• Viewing rectangle / rectangle of influence.

What we have made

The red cars are minimum aggressive level cars of class 1.

Green car is changing the lane according to heuristic what we have defined depicting the class 3

Here the concept of queuing is depicted as most of the cars with same heuristic is moving in a queue.

The concept of flocking is depicted in this image where the central 5 red cars are moving in a flock.

All the cars exit from the right side of the screen.

Future ScopeFuture Scope• To implement in driving simulators for learning

purposes.• It can be used in games and movies to provide

visual effects.• Game Development• Machine Learning• Intelligent Toys

ReferencesReferences• http://www.riemers.net• http://www.gamedev.net/reference• http://www.alinenormoyle.com/projects.html• Steering Behaviors For Autonomous Characters• Programming Game AI by Example - Mat Buckland• http://sandbox.siggraph.org/papers.html • http://sketchup.google.com/ • http://www.youtubekeep.com/download-video/india-driving/153521

• N. Pelechano, J. M. Allbeck, and N. I. Badler.

Controlling individual agents in high-density crowdsimulation. In SCA ’07: Proceedings of the 2007 ACMSIGGRAPH/Eurographics symposium on Computeranimation, pages 99–108, 2007.

top related