How to start PSM_NODE (PSM Scene Recognition + Pose Prediction):

How to start psm_node:
roscore
start kinematic chain - roslaunch asr_kinematic_chain_dome transformation_publisher_left.launch
start the server - roslaunch asr_recognizer_prediction_psm psm_node
call the service - roscd asr_recognizer_prediction_psm
python python/psm_node_test.py


Note:

There are no hypothesis generated for the background scene. See the python/psm_node_test.py for an example usage.

What are the parameters of the service call:


asr_msgs/AsrObject[] objects                - input: the list of all oberved objects
---
next_best_view/AttributedPointCloud pose_hypothesis - output: the generated hypothesis












How to start ONLY pose prediciton:
1. start kinematic chain (needed for transformation into target frame eg. /map) - roslaunch kinematic_chain transformation_publishers_left.launch
2. it seems that the kinematic_chain does not contain all transformations. Start asr_flir_ptu_driver - roslaunch asr_flir_ptu_driver ptu_left.launch
3. rosrun asr_recognizer_prediction_psm recognizer_prediction_psm
4. python pyhton/recognizer_prediction_test.py or rosrun asr_recognizer_prediction_psm recognizer_prediction_psm_client 10


What are the parameters of the service call:

string path                     - input: the path to the xml file that contains the scenes (eg. models/breakfast.xml)
string[] scenes                     - input: the name of the scenes. It gerneates hypothesis only for these scenes. 
float32[] scene_probabilities               - input: contains the probability of the scene
uint32 num_votes                    - input: the overall number of hypothesis that should be generated
asr_msgs/AsrObject[] objects                - input: the list of all oberved objects
string base_frame_id                    - input: the base frame to which the observed objects and the hypothesis should be transformed into
---
next_best_view/AttributedPointCloud pose_hypothesis - output: the generated hypothesis



Note:
- the number of scene_probabilities has to be equal to the number of scenes
- the sum of the scene_probabilities should be 1.0, otherwise the number of hypothesis will not be correct
- there has to be at least one observed object
- if there are more than one object observed, the first object in the objects list that occurs in the scene will be treated as reference object


See 
http://i61p109.ira.uka.de/twiki/bin/view/IlcasProjects/PosePredictionPsm
for further documentation.