Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
doc:reasoning_about_objects [2014/11/28 14:56]
admin
doc:reasoning_about_objects [2016/04/19 08:11] (current)
daniel86 [Visualize the environment, including CAD models of objects]
Line 1: Line 1:
 ====== Reasoning about objects ====== ====== Reasoning about objects ======
 +~~NOTOC~~
 +\\
 +^ This page describes the '​catkinized'​ version of KnowRob that uses the [[http://​wiki.ros.org/​catkin/​|catkin buildsystem]] and the pure Java-based [[http://​wiki.ros.org/​rosjava|rosjava]]. The documentation for the older version, which was based on the rosbuild buildsystem and rosjava_jni,​ can be found [[/​doc/​reasoning_about_objects?​rev=1407309271|here]].^
 +\\
  
 Before starting this tutorial, you should have completed the following ones: Before starting this tutorial, you should have completed the following ones:
Line 10: Line 14:
 With the following commands, you can launch KnowRob including the semantic map of our laboratory kitchen, and then load the OWL file containing a set of demo objects. You need to have the knowrob_tutorials repository in your catkin workspace. With the following commands, you can launch KnowRob including the semantic map of our laboratory kitchen, and then load the OWL file containing a set of demo objects. You need to have the knowrob_tutorials repository in your catkin workspace.
 <code prolog> <code prolog>
- rosrun rosprolog rosprolog knowrob_basics_tutorial +?- rosrun rosprolog rosprolog knowrob_basics_tutorial 
- ​owl_parse('​package://​knowrob_map_data/​owl/​ccrl2_semantic_map.owl'​). +?- owl_parse('​package://​knowrob_map_data/​owl/​ccrl2_semantic_map.owl'​). 
- ​owl_parse('​package://​knowrob_basics_tutorial/​owl/​ccrl2_map_objects.owl'​).+?- owl_parse('​package://​knowrob_basics_tutorial/​owl/​ccrl2_map_objects.owl'​).
 </​code>​ </​code>​
  
Line 21: Line 25:
  
 Once the map and the example objects have been loaded, you can start the visualization canvas, [[http://​localhost:​1111|open the web-based visualization in your browser]], and add the objects to this 3D canvas. If definitions for CAD models of objects are loaded, these models are automatically used for visualizing object instances. Otherwise, the system uses grey boxes as default visualization models. Once the map and the example objects have been loaded, you can start the visualization canvas, [[http://​localhost:​1111|open the web-based visualization in your browser]], and add the objects to this 3D canvas. If definitions for CAD models of objects are loaded, these models are automatically used for visualizing object instances. Otherwise, the system uses grey boxes as default visualization models.
 +
 +Please make sure that you have launched the '​rosbridge'​ server in another terminal using:
 +<code bash>
 +roslaunch rosbridge_server rosbridge_websocket.launch
 +</​code>​
 +
 +
 +You can then send the following commands from your KnowRob shell:
 <code prolog> <code prolog>
-?- visualisation_canvas.+?- visualisation_server.
  
-open your browser at http://​localhost:​1111+Ignore the INFO output on the console 
 +% Open your browser at http://​localhost:​1111
  
-?- owl_individual_of(Map,​ knowrob:'​SemanticEnvironmentMap'​), ​add_object_with_children(Map).+?- owl_individual_of(Map,​ knowrob:'​SemanticEnvironmentMap'​), ​marker_update(object(Map)).
 </​code>​ </​code>​
 Similarly, we can also visualize small objects in the kitchen: Similarly, we can also visualize small objects in the kitchen:
Line 33: Line 46:
   owl_individual_of(A,​ knowrob:'​FoodVessel'​);​   owl_individual_of(A,​ knowrob:'​FoodVessel'​);​
  ​owl_individual_of(A,​ knowrob:'​FoodOrDrink'​)),​  ​owl_individual_of(A,​ knowrob:'​FoodOrDrink'​)),​
- add_object(A, _).+ marker_update(object(A)).
 </​code>​ </​code>​
  
Line 41: Line 54:
 Since all objects in the map are instances of the respective object classes, one can query for objects that have certain properties or belong to a certain class, for example: Since all objects in the map are instances of the respective object classes, one can query for objects that have certain properties or belong to a certain class, for example:
 <code prolog> <code prolog>
- % perishable objects: +% perishable objects: 
- ?- owl_individual_of(A,​ knowrob:'​Perishable'​). +?- owl_individual_of(A,​ knowrob:'​Perishable'​). 
- A = knowrob:'​butter1'​ ; +A = map_obj:'​butter1'​ ; 
- A = knowrob:'​buttermilk1'​ ; +A = map_obj:'​buttermilk1'​ ; 
- A = knowrob:'​cheese1'​ ; +A = map_obj:'​cheese1'​ ; 
- A = knowrob:'​milk1'​ ; +A = map_obj:'​milk1'​ ; 
- A = knowrob:'​yogurt1'​ ; +A = map_obj:'​yogurt1'​ ; 
- A = knowrob:'​sausage1'​+A = map_obj:'​sausage1'​ 
 + 
 +% all HandTools (e.g. silverware) 
 +?- owl_individual_of(A, ​knowrob:'​HandTool'​). 
 +A = map_obj:'​knife1'​ ; 
 +A = map_obj:'​fork1'​ 
 + 
 +% all FoodVessels (i.e. pieces of tableware) 
 +?- owl_individual_of(A,​ knowrob:'​FoodVessel'​). 
 +A = map_obj:'​cup1'​ ; 
 +A = map_obj:'​plate1'​ ; 
 +A = map_obj:'​saucer1'​ 
 + 
 +% everything with a handle: 
 +?- owl_has(A, knowrob:​properPhysicalParts,​ H),  
 +   ​owl_individual_of(H, ​ knowrob:'​Handle'​). 
 +A = map_obj:'​Dishwasher37',​ 
 +H = map_obj:'​Handle145'​ 
 +[...] 
 +</​code>​ 
 + 
 + 
 +===== Querying for qualitative spatial relations ===== 
 + 
 +Using [[reason_using_computables|computables]] that calculate qualitative spatial relations between objects, we can query e.g. in which container we expect to find sausage1, ask for the content of Refrigerator67,​ or ask what is on top of Dishwasher37:​ 
 + 
 +<code prolog>​ 
 + ?- rdf_triple(knowrob:'​in-ContGeneric',​ map_obj:​sausage1,​ C). 
 + C = knowrob:'​Refrigerator67'
    
- % all HandTools (e.g. silverware) + ?​- ​rdf_triple(knowrob:'​in-ContGeneric',​ O, knowrob:'​Refrigerator67'). 
- ?- owl_individual_of(A, knowrob:'​HandTool'). + map_obj:'cheese1' ; 
- knowrob:'knife1' ; + map_obj:'​milk1'​ ; 
- knowrob:'fork1'+ O = map_obj:'sausage1'
    
- % all FoodVessels (i.e. pieces of tableware) + ?​- ​rdf_triple(knowrob:'​on-Physical'A, knowrob:'​Dishwasher37'). 
- ?- owl_individual_of(A, knowrob:'​FoodVessel'). + A = map_obj:'cup1' ​; 
- A = knowrob:'​cup1'​ ; + = knowrob:'​CounterTop205'
- A = knowrob:'​plate1'​ ; +
- A = knowrob:'​saucer1'​ +
-  +
- % everything with a handle: +
- ?- owl_has(A, knowrob:​properPhysicalParts,​ H),  +
-    owl_individual_of(H,  ​knowrob:'​Handle'). +
- A = knowrob:'Dishwasher37', +
- = knowrob:'​Handle145'+
 </​code>​ </​code>​
  
  
- +===== Inferring ​likely storage locations =====
-===== Infer likely storage locations =====+
  
 For some tasks, robots need to reason about the nominal locations of objects, for example when cleaning up or when unpacking a shopping basket. There are different techniques for inferring the location where an object should be placed. ​ For some tasks, robots need to reason about the nominal locations of objects, for example when cleaning up or when unpacking a shopping basket. There are different techniques for inferring the location where an object should be placed. ​
Line 78: Line 110:
  
  
-=== Query for likely storage ​location ​===+=== Querying ​for likely storage ​locations ​===
  
 The simple option based on the storagePlaceFor predicate can be queried as follows in order to determine where an object (instance or class) shall be stored, or which known objects are to be stored in a given container: The simple option based on the storagePlaceFor predicate can be queried as follows in order to determine where an object (instance or class) shall be stored, or which known objects are to be stored in a given container:
 <code prolog> <code prolog>
- ?- storagePlaceFor(Place,​ map_obj:'​butter1'​). +?- storagePlaceFor(Place,​ map_obj:'​butter1'​). 
- Place = knowrob:'​Refrigerator67'​+Place = knowrob:'​Refrigerator67'​
  
- ?- storagePlaceFor(knowrob:'​Refrigerator67',​ Obj). +?- storagePlaceFor(knowrob:'​Refrigerator67',​ Obj). 
- Obj = knowrob:'​butter1'​ ; +Obj = map_obj:'​butter1'​ ; 
- Obj = knowrob:'​buttermilk1'​ ; +Obj = map_obj:'​buttermilk1'​ ; 
- Obj = knowrob:'​cheese1'​ ; +Obj = map_obj:'​cheese1'​ ; 
- Obj = knowrob:'​milk1';​ +Obj = map_obj:'​milk1';​ 
- [...]+[...]
 </​code>​ </​code>​
  
Line 96: Line 128:
 <code prolog> <code prolog>
  ?- storagePlaceForBecause(Place,​ map_obj:'​butter1',​ Because).  ?- storagePlaceForBecause(Place,​ map_obj:'​butter1',​ Because).
- Place = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Refrigerator67',​ + Place = knowrob:'Refrigerator67',​ 
- ​Because = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Perishable'​+ ​Because = knowrob:'Perishable'​
 </​code>​ </​code>​
  
Line 105: Line 137:
  
  
-=== Check for objects that are not at their storage location ===+=== Checking ​for objects that are not at their storage location ===
  
-By combining the semantic map, giving ​the current object locationswith the knowledge about the locations where objects should be, the robot can determine which objects are mis-placed. This may for instance be useful when tidying up an environment. +By combining the semantic map and the current object locations with the knowledge about the locations where objects should be, the robot can determine which objects are misplaced. This may for instance be useful when tidying up an environment. The following code first compute which objects are inside which other ones, then selects those that are food or drink, computes their most likely storage place (which, in this example, is usually the refrigerator),​ and compares the two places. As a result, the system finds out that the butter and the buttermilk are misplaced in a drawer and in the oven, respectively.
- +
-The following code first compute which objects are inside which other ones, then selects those that are food or drink, computes their most likely storage place (which, in this example, is usually the refrigerator),​ and compares the two places. As a result, the system finds out that the butter and the buttermilk are misplaced in a drawer and in the oven, respectively.+
  
 <code prolog> <code prolog>
Line 116: Line 146:
     storagePlaceFor(StoragePlace,​Obj), ​     storagePlaceFor(StoragePlace,​Obj), ​
     ActualPlace\=StoragePlace.     ActualPlace\=StoragePlace.
- Obj = 'http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​butter1',​ + Obj = map_obj:'butter1',​ 
- ​ActualPlace = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Drawer31',​ + ​ActualPlace = knowrob:'Drawer31',​ 
- ​StoragePlace = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Refrigerator67'​ ; + ​StoragePlace = knowrob:'Refrigerator67'​ ; 
- Obj = 'http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​buttermilk1',​ + Obj = map_obj:'buttermilk1',​ 
- ​ActualPlace = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Oven19',​ + ​ActualPlace = knowrob:'Oven19',​ 
- ​StoragePlace = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Refrigerator67'​+ ​StoragePlace = knowrob:'Refrigerator67'​
 </​code>​ </​code>​
  
-===== Read object component hierarchy =====+You will notice that it takes a bit of time to answer the query: This is because the first predicate searches for all objects that are inside another one, and for doing this has to compute all pairwise '​inside'​ relations between all objects. By restricting the set of objects to be considered as inner and outer objects, this could be sped up a lot. This is one example of a general property of Prolog: The efficiency depends on how the query is formulated.
  
-Composed objects and their parts are linked by an inverse part-of ​hierarchy, described using the properPhysicalParts property in OWL. This property is transitive, i.e. a part of a part of an object is also a part of the object itself. You can read all parts and sub-parts of an object using owl_has, which takes the transitivity into account. In the example below, Handle160 is a part of Door70, which by itself is part of Refrigerator67.+ 
 +===== Reading the object component hierarchy ===== 
 + 
 +Composed objects and their parts are linked by a '​parts' ​hierarchy, described using the properPhysicalParts property in OWL. This property is transitive, i.e. a part of a part of an object is also a part of the object itself. You can read all parts and sub-parts of an object using owl_has, which takes the transitivity into account. In the example below, Handle160 is a part of Door70, which by itself is part of Refrigerator67.
 <code prolog> <code prolog>
  ?- owl_has(knowrob:'​Refrigerator67',​ knowrob:​properPhysicalParts,​ P).  ?- owl_has(knowrob:'​Refrigerator67',​ knowrob:​properPhysicalParts,​ P).
- P = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Door70'​ ; + P = knowrob:'Door70'​ ; 
- P = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Handle160'​ ; + P = knowrob:'Handle160'​ ; 
- P = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#Hinge70'​+ P = knowrob:'Hinge70'​
 </​code>​ </​code>​
  
-=== Special case: articulated objects such as cupboards === 
  
-{{ :part-of-hierarchy-map.png?​nolink&​472| }}+===== Representation ​of articulated objects such as cupboards =====
  
-Articulated objects like cupboards that have doors or drawers are represented in a special way to describe, on the one hand, the component hierarchy, and on the other hand the fixed and moving connections. Like for other composed objects, there is a part-of hierarchy ​(properPhysicalParts). Joints are parts of the cupboard/​parent object, and are fixed-to (connectedTo-Rigidly) both the parent and the child (e.g. the door). In addition, the child is hingedTo or prismaticallyConnectedTo the parent.+{{ :part-of-hierarchy-map.png?​nolink&​472| }}
  
 +Articulated objects, e.g. cupboards, that have doors or drawers are represented in a special way to describe, on the one hand, the component hierarchy, and, on the other hand, which connections are fixed and which are movable. Like for other composed objects, there is a part-of hierarchy (properPhysicalParts). Joints are parts of the cupboard/​parent object, and are fixed-to (connectedTo-Rigidly) both the parent and the child (e.g. the door). In addition, the child is hingedTo or prismaticallyConnectedTo the parent.
  
 Joints are described using the following properties, which are compatible to the representation used by the [http://​www.ros.org/​wiki/​articulation ROS articulation stack]. Joints are described using the following properties, which are compatible to the representation used by the [http://​www.ros.org/​wiki/​articulation ROS articulation stack].
Line 150: Line 183:
  
  
-===== Read articulation information =====+===== Reading ​articulation information =====
  
-There are some convenience ​predicates for reading, creating, updating and deleting joints from articulated objects. This task is on the one hand rather common, on the other hand somewhat complex because the structure visualized in the previous image needs to be established. +There are some helper ​predicates for reading, creating, updating and deleting joints from articulated objects. This task is on the one hand rather common, on the other hand somewhat complex because the structure visualized in the previous image needs to be established. To create a joint of type knowrob:'​HingedJoint'​ between two object instances roboearth:'​IkeaExpedit2x40'​ and roboearth:'​IkeaExpeditDoor13'​ at position (1,1,1) with unit orientation,​ radius 0.33m and joint limits 0.1 and 0.5 respectively,​ one can use the following statement:
- +
-To create a joint of type knowrob:'​HingedJoint'​ between two object instances roboearth:'​IkeaExpedit2x40'​ and roboearth:'​IkeaExpeditDoor13'​ at position (1,1,1) with unit orientation,​ radius 0.33m and joint limits 0.1 and 0.5 respectively,​ one can use the following statement:+
 <code prolog> <code prolog>
  ​create_joint_information('​HingedJoint',​ roboearth:'​IkeaExpedit2x40',​ roboearth:'​IkeaExpeditDoor13', ​  ​create_joint_information('​HingedJoint',​ roboearth:'​IkeaExpedit2x40',​ roboearth:'​IkeaExpeditDoor13', ​
Line 160: Line 191:
 </​code>​ </​code>​
  
-If a prismatic joint is to be created instead, the empty list [] needs to be replaced with a unit vector describing the joint'​s direction, e.g. [0,0,1] for a joint opening in z-direction,​ and the joint type needs to be set as '​PrismaticJoint'​. +If a prismatic joint is to be created instead, the empty list [] needs to be replaced with a unit vector describing the joint'​s direction, e.g. [0,0,1] for a joint opening in z-direction,​ and the joint type needs to be set as '​PrismaticJoint'​. Joint information can conveniently be read using the following predicate that requires a joint instance as argument:
- +
-Joint information can conveniently be read using the following predicate that requires a joint instance as argument:+
 <code prolog> <code prolog>
  ​read_joint_information(Joint,​ Type, Parent, Child, Pose, Direction, Radius, Qmin, Qmax).  ​read_joint_information(Joint,​ Type, Parent, Child, Pose, Direction, Radius, Qmin, Qmax).
Line 172: Line 201:
 </​code>​ </​code>​
  
-===== Read and convert units of measure ===== 
  
-All numerical values can optionally be annotated with a unit of measure. To keep the system backwards compatible, other values are interpreted to be given in the respective SI unit (e.g. meter, second). 
  
-Full article incl. explanation of design choices and links to further information:​ [[Measurement_units]] +===== Reading and converting ​units of measure =====
-<code prolog>​ +
- $ roscd knowrob_common/​owl +
- $ rosrun rosprolog rosprolog ias_knowledge_base +
-  +
- ?- owl_parse('​knowrob_units.owl',​ false, false, true). +
- ?- consult('​../​prolog/​knowrob_units.pl'​). +
-  +
- % read information that is asserted for a test instance +
- ?- rdf_has('​http://​ias.cs.tum.edu/​kb/​knowrob_units.owl#​test-inst',​ +
-            '​http://​ias.cs.tum.edu/​kb/​knowrob_units.owl#​length',​ O). +
- ​O ​literal(type('​http://​qudt.org/​vocab/​unit#​Centimeter','​12.0'​)) . +
-  +
- % manual conversion into other units +
- ?- convert_to_unit($O,​ '​http://​qudt.org/​vocab/​unit#​Kilometer',​ P). +
- ​P ​0.00012. +
-  +
- ?- convert_to_unit($O,​ '​http://​qudt.org/​vocab/​unit#​Meter',​ P). +
- ​P ​0.12. +
-  +
- ?- convert_to_unit($O,​ '​http://​qudt.org/​vocab/​unit#​Millimeter',​ P). +
- ​P ​120.0. +
-</​code>​+
  
-The integration ​with the rdf_triple computables allows to transparently convert values into the desired ​unit of measure:  +All numerical values in KnowRob can optionally be annotated ​with unit of measure. To keep the system backwards compatible, ​ values without annotation are interpreted to be given in the respective SI unit (e.gmetersecond). The full article including an explanation of the design choices and links to further information can be found [[/doc/measurement_units|here]].
-<code prolog>​ +
- % transparent conversion during ​the query   +
- ?- rdf_triple('​http://​ias.cs.tum.edu/​kb/​knowrob_units.owl#​length'​ +
-               '​http://​ias.cs.tum.edu/​kb/​knowrob_units.owl#​test-inst',​  +
-                literal(type('​http://​qudt.org/​vocab/​unit#​Meter',​ Val))). +
- Val = 0.12 ; +
-  +
- ?- rdf_triple('​http:​//ias.cs.tum.edu/​kb/​knowrob_units.owl#​length',​  +
-               '​http://​ias.cs.tum.edu/​kb/​knowrob_units.owl#​test-inst',​  +
-                literal(type('​http://​qudt.org/​vocab/​unit#​Kilometer',​ Val))). +
- Val = 0.00012 ; +
-</​code>​+
  
- 
-===== Query for qualitative spatial relations ===== 
- 
-Using [[reason_using_computables|computables]] that calculate qualitative spatial relations between objects, we can query e.g. in which container we expect to find sausage1, ask for the content of Refrigerator67,​ or ask what is on top of Dishwasher37:​ 
- 
-<code prolog> 
- ?- rdf_triple(knowrob:'​in-ContGeneric',​ map_obj:​sausage1,​ C). 
- C = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#​Refrigerator67'​ 
-  
- ?- rdf_triple(knowrob:'​in-ContGeneric',​ O, knowrob:'​Refrigerator67'​). 
- O = '​http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​cheese1'​ ; 
- O = '​http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​milk1'​ ; 
- O = '​http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​sausage1'​ 
-  
- ​rdf_triple(knowrob:'​on-Physical',​ A, knowrob:'​Dishwasher37'​). 
- A = '​http://​ias.cs.tum.edu/​kb/​ccrl2_map_objects.owl#​cup1'​ ; 
- A = '​http://​ias.cs.tum.edu/​kb/​knowrob.owl#​CounterTop205'​ 
-</​code>​