Skip to content

Java implementation of the ongoing research in motivation to be integrated in cafer

Notifications You must be signed in to change notification settings

doncieux/mdb_motiven

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Motiven in MDB

Java implementation of the ongoing research in motivation to be integrated in cafer

Libs installation

To use this project, it is needed to install three libraries in the maven repository. This libraries are in folder lib, and for its instalation it must be executed:

$ mvn install:install-file -Dfile=lib/artificial_neural_networks-1.01.jar -DgroupId=es.udc.gii.common -DartifactId=evolutionary_algorithms -Dversion=1.0 -Dpackaging=jar
$ mvn install:install-file -Dfile=lib/evolutionary_algorithms-1.0.jar -DgroupId=es.udc.gii.common -DartifactId=artificial_neural_networks -Dversion=1.01 -Dpackaging=jar
$ mvn install:install-file -Dfile=lib/anji-1.0.jar -DgroupId=com -DartifactId=anji -Dversion=1.0 -Dpackaging=jar

Description

This implementation is based on creating “certainty areas” that represent both areas with a concrete goal (or subgoal) and the subgoal of the next area. These areas can expand and shrink dynamically and individually for any of the dimensions of the space state (or perceptions). To determine which points of the state space will constitute the certainty area, the correlation of perceptual state traces will be studied. Perceptions for which a large correlation is found are used to determine which points should belong to the certainty area and participate in the evaluation of candidate actions (as a value function) to be chosen to reach the goal.

USAGE

Important: Due to an error in anji-neat library, it is mandatory create a folder to use neat, what is set up in the neat properties file (it is already created in this program, see config folder).

The MDB is thought as an independent cognitive architecture to be run with any robot or simulator. For this, to run the MDB, it is needed to run the simulator first, so the MDB will connect to it by TCP/IP. After running the rimulator or robot, the MDB can be executed. As this is an ongoing unstable version of MDB, some code makes reference to the configuration of the current test experiment so, for now, it only will work with the simulator provided in Collect a ball java simulator.

To run MDB:

$ java -jar MDBCore-full config/MDB-Config.xml

About

Java implementation of the ongoing research in motivation to be integrated in cafer

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages