knowrob: bosch_semantic_map | comp_cop | comp_germandeli | comp_orgprinciples | comp_spatial | comp_temporal | ias_knowledge_base | ias_prolog_addons | ias_semantic_map | jpl | json_prolog | knowrob_actions | knowrob_cad_models | knowrob_common | knowrob_objects | mod_probcog | mod_srdl | mod_vis | rosprolog | semweb | srldb | tf_prolog | thea

Package Summary

Implementation of the Semantic Robot Description Language (SRDL) that provides model and reasoning tools to match prerequisites of actions to the capabilities of a robot.

• Author: Moritz Tenorth, Tobias Roehm
• License: GPL
• Repository: tum-ros-pkg
• Source: svn http://code.in.tum.de/pubsvn/knowrob/tags/latest/mod_srdl

New version available

A complete re-implementation of SRDL, srdl2, is available and contained in the same ROS package mod_srdl. Please have a look at the tutorial and documentation here: http://ias.in.tum.de/kb/wiki/index.php/SRDL2_Tutorial

Old version

Getting Started

The SRDL functionality is packaged in the ROS package mod_srdl. In order to use it follow these steps:

1. Make sure to have a working copy of package mod_srdl on your computer and run

rosmake mod_srdl

2. Start srdl with the command

rosrun rosprolog rosprolog mod_srdl

Some basic queries

Here are examples for some basic queries to get started:

All subactions of an action:

 subActions(srdl_action:'TableSetting', L).

All components of a robot:

 hasComponent(rosie:'TUM_Rosie_RobotInstance1', C).

Match based on robot components

 matchRobotAndAction(rosie:'TUM_Rosie_RobotInstance1', srdl_action:'TableSetting').
 matchRobotAndAction(rosie:'TUM_Rosie_RobotInstance1', srdl_action:'ChocolateCakeBaking').

All missing capabilities for an unfeasible action:

 returnMissingCapsForAction(srdl_action:'ChocolateCakeBaking',  rosie:'TUM_Rosie_RobotInstance1', Caps).

Print all components that are missing for an unfeasible action (order via capability and capability provision alternative):

 printMissingComponents(srdl_action:'ChocolateCakeBaking', rosie:'TUM_Rosie_RobotInstance1').

Compute success probability (based on experience) for an action:

 computeSuccessProbability(srdl_action:'TableSetting', P).

Show technical specification/ attributes of components (as far as specified/ contained in KB):

 printAttributesOfComponent(pr2:'PR2_l_ArmComposition').

Check if a capability can be learned:

 verifyCapAvailability(srdl_cap:'RedCupRecognitionCapability', rosie:'TUM_Rosie_RobotInstance1').
 isLearnableCapability(srdl_cap:'RedCupRecognitionCapability', rosie:'TUM_Rosie_RobotInstance1').

Actions in which TUM Rosie and PR2 differ

Example 1 - Driving backwards (possible for Rosie but not for PR2):

 matchRobotAndAction(rosie:'TUM_Rosie_RobotInstance1', srdl_action:'DriveBackwards').
 matchRobotAndAction(pr2:'PR2_Instance1', srdl_action:'DriveBackwards').
 returnMissingCapsForAction(srdl_action:'DriveBackwards',  pr2:'PR2_Instance1', Caps).
 printMissingComponents(srdl_action:'DriveBackwards', pr2:'PR2_Instance1').

Example 2 - Crushing a can of coke using strong grip (possible for PR2 but not for Rosie):

 matchRobotAndAction(rosie:'TUM_Rosie_RobotInstance1', srdl_action:'CrushCokeCan').
 returnMissingCapsForAction(srdl_action:'CrushCokeCan',  rosie:'TUM_Rosie_RobotInstance1', Caps).
 printMissingComponents(srdl_action:'CrushCokeCan', rosie:'TUM_Rosie_RobotInstance1').
 matchRobotAndAction(pr2:'PR2_Instance1', srdl_action:'CrushCokeCan').