wrap initialization for anymal

script/scenarios/sandbox/anymal_one_step/anymal_circle_one_step.py

@@ -188,35 +188,51 @@ if (not cg_success) or cg_too_many_states or (not cg_reach_goal):
 
 
 import hpp.corbaserver.rbprm.fewstepsplanner as sp
+import time
+def dispContactPlan(states, step = 0.5):
+	for s in states:
+		v(s.q());
+		time.sleep(step)
 
-states = sp.interpolateState(fullBody, 0.002,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True)
 
 fsp = sp.FewStepPlanner(tp.cl,tp.ps,tp.rbprmBuilder, fullBody, pathPlayer = tp.pp)
+print "contact start "
+states, cfgs = fsp.interpolateStates(0.002,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True)
+print "contact start "
 
 
 
 #~ print "names ", fsp.rbprmBuilder.getAllJointNames()
 #~ print "names ", tp.rbprmBuilder.getAllJointNames()
 
-n_goal = tp.q_goal[:7]
-n_goal[0] += 1
+n_goal = tp.q_goal[:7][:]
+n_goal[0] += 2
+n_goal[1] += 1
+n_goal[3:7] = [0.,0.,0.7071,0.7071]
 
 
 
-goal_state = states[-1].q()[:]
-goal_state[:7] = n_goal
-fullBody.setStartState(states[-1].q(),[fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
-fullBody.setEndState(goal_state, [fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
+n_goal_state = states[-1].q()[:]
+n_goal_state[:7] = n_goal[:]
+fullBody.setStartStateId(states[-1].sId)
+fullBody.setEndState(n_goal_state, [fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
 
 pId= fsp.guidePath(tp.q_goal[:7],n_goal)
-states2 = sp.interpolateState(fullBody, 0.002,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True, erasePreviousStates = False)
+states2, cfgs2 = fsp.interpolateStates(0.002,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True, erasePreviousStates = False)
 
+#~ displayContactSequence(v,cfgs2,0.1)
+print "cplan"
+#~ dispContactPlan(states2,0.1)
 
-goal_state[6] = -1
+print "end cplan"
 
-fullBody.setStartState(states[-1].q(),[fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
-fullBody.setEndState(goal_state, [fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
+pId= fsp.guidePath(n_goal, tp.q_goal[:7])
+fullBody.setStartState(n_goal_state,[fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
+fullBody.setEndState(states[-1].q(), [fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],normals)
 
-pId= fsp.guidePath(tp.q_goal[:7],n_goal)
-states3 = sp.interpolateState(fullBody, 0.001,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True, erasePreviousStates = False)
+print "new state cplan"
+#~ dispContactPlan(states2,0.1)
+states3, cfgs3 = fsp.interpolateStates(0.002,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True, erasePreviousStates = False)
 
+n_goal[1] -= 3
+states4, cfgs4 = fsp.goToQuasiStatic(states3[-1],n_goal)

script/scenarios/sandbox/anymal_one_step/config.py

+SCENE="multicontact/ground"
+INIT_CONFIG_ROOT =   [0,0,0.465,0,0,0,1]
+INIT_CONFIG_WB   =   None
+ROOT_BOUNDS      =  [-20,20, -20, 20, 0.4, 0.5]

script/scenarios/sandbox/anymal_one_step/demo.py

+import setup_one_step as sos
+
+
+fsp = sos.fewStepPlanner #planner instance
+q_init =  sos.q_init  #initial configuration
+initState = sos.initState #initial state. 
+viewer = sos.v
+
+### Go somewhere
+
+n_goal = q_init[:7][:]
+n_goal[0] += 2
+n_goal[1] += 1
+n_goal[3:7] = [0.,0.,0.7071,0.7071]
+states, configs = fsp.goToQuasiStatic(initState,n_goal, displayGuidePath = True)
+#equivalent to
+ # fsp.goToQuasiStatic(self, initState, n_goal, stepsize = 0.002, goalLimbsInContact = None, goalNormals = None, displayGuidePath = False)
+
+
+#display computed States:
+sos.dispContactPlan(states,0.051) #2nd argument is frame rateue
+
+
+s = states[-1] #last state computed
+
+#display configuration
+viewer(s.q())
+
+#some helpers: 
+s.q() # configuration associated to state
+#~ initState. # configuration associated to state

script/scenarios/sandbox/anymal_one_step/setup_one_step.py

+import config as cfg
+
+
+#root path configuration
+
+from hpp.corbaserver.rbprm.anymal_abstract import Robot
+from hpp.gepetto import Viewer
+from hpp.corbaserver import ProblemSolver
+import numpy as np
+import time
+statusFilename = "/tmp/infos.log"
+
+
+vMax = 0.3# linear velocity bound for the root
+aMax = 1.# linear acceleration bound for the root
+extraDof = 6
+mu=0.5# coefficient of friction
+# Creating an instance of the helper class, and loading the robot
+# Creating an instance of the helper class, and loading the robot
+rbprmBuilder = Robot ()
+# Define bounds for the root : bounding box of the scenario
+root_bounds =  cfg.ROOT_BOUNDS
+rbprmBuilder.setJointBounds ("root_joint", root_bounds)
+
+# The following lines set constraint on the valid configurations:
+# a configuration is valid only if all limbs can create a contact with the corresponding afforcances type
+rbprmBuilder.setFilter(rbprmBuilder.urdfNameRom)
+for rom in rbprmBuilder.urdfNameRom :
+    rbprmBuilder.setAffordanceFilter(rom, ['Support'])
+
+# We also bound the rotations of the torso. (z, y, x)
+#~ rbprmBuilder.boundSO3([-1.7,1.7,-0.1,0.1,-0.1,0.1])
+rbprmBuilder.boundSO3([-3,3,-3.,3.,-3.,-3])
+# Add 6 extraDOF to the problem, used to store the linear velocity and acceleration of the root
+rbprmBuilder.client.robot.setDimensionExtraConfigSpace(extraDof)
+# We set the bounds of this extraDof with velocity and acceleration bounds (expect on z axis)
+rbprmBuilder.client.robot.setExtraConfigSpaceBounds([-vMax,vMax,-vMax,vMax,0,0,-aMax,aMax,-aMax,aMax,0,0])
+indexECS = rbprmBuilder.getConfigSize() - rbprmBuilder.client.robot.getDimensionExtraConfigSpace()
+
+# Creating an instance of HPP problem solver 
+psGuide = ProblemSolver( rbprmBuilder )
+# define parameters used by various methods : 
+psGuide.setParameter("Kinodynamic/velocityBound",vMax)
+psGuide.setParameter("Kinodynamic/accelerationBound",aMax)
+psGuide.setParameter("DynamicPlanner/sizeFootX",0.01)
+psGuide.setParameter("DynamicPlanner/sizeFootY",0.01)
+psGuide.setParameter("DynamicPlanner/friction",mu)
+# sample only configuration with null velocity and acceleration :
+psGuide.setParameter("ConfigurationShooter/sampleExtraDOF",False)
+
+# initialize the viewer :
+from hpp.gepetto import ViewerFactory
+vf = ViewerFactory (psGuide)
+
+# load the module to analyse the environnement and compute the possible contact surfaces
+from hpp.corbaserver.affordance.affordance import AffordanceTool
+afftool = AffordanceTool ()
+afftool.setAffordanceConfig('Support', [0.5, 0.03, 0.00005])
+afftool.loadObstacleModel ("hpp_environments", cfg.SCENE, "planning", vf)
+try :
+    v = vf.createViewer(displayArrows = True)
+except Exception:
+    print "No viewer started !"
+    class FakeViewer():
+        def __init__(self):
+            return
+        def __call__(self,q):
+            return
+    v = FakeViewer()
+    
+#afftool.visualiseAffordances('Support', v, v.color.lightBrown)
+
+q_init = rbprmBuilder.getCurrentConfig ();
+q_init[:7] = cfg.INIT_CONFIG_ROOT
+
+
+# Choosing RBPRM shooter and path validation methods.
+psGuide.selectConfigurationShooter("RbprmShooter")
+psGuide.selectPathValidation("RbprmPathValidation",0.05)
+# Choosing kinodynamic methods :
+psGuide.selectSteeringMethod("RBPRMKinodynamic")
+psGuide.selectDistance("Kinodynamic")
+psGuide.selectPathPlanner("DynamicPlanner")
+
+# Solve the planning problem :
+# move the robot out of the view before computing the contacts
+
+from hpp.corbaserver import Client
+ #~ #DEMO code to play root path and final contact plan
+cl = Client()
+cl.problem.selectProblem("rbprm_path")
+rbprmBuilder2 = Robot ("toto")
+ps2 = ProblemSolver( rbprmBuilder2 )
+cl.problem.selectProblem("default")
+#~ cl.problem.movePathToProblem(pId,"rbprm_path",rbprmBuilder.getAllJointNames()[1:])
+r2 = Viewer (ps2, viewerClient=v.client)
+q_far = q_init[::]
+q_far[2] = -2
+r2(q_far)
+
+from hpp.gepetto import PathPlayer
+pp = PathPlayer (v)
+
+###### WHOLE BODY INIT
+
+from hpp.corbaserver.rbprm.anymal import Robot
+#~ from hpp.gepetto import Viewer
+from tools.display_tools import *
+import time
+
+pId = 0
+
+fullBody = Robot ()
+# Set the bounds for the root
+fullBody.setJointBounds ("root_joint",  root_bounds)
+## reduce bounds on joints along x, to put conservative condition on the contact generation for sideway steps
+fullBody.setVeryConstrainedJointsBounds()
+# add the 6 extraDof for velocity and acceleration (see *_path.py script)
+fullBody.client.robot.setDimensionExtraConfigSpace(extraDof)
+fullBody.client.robot.setExtraConfigSpaceBounds([-vMax,vMax,-vMax,vMax,0,0,-aMax,aMax,-aMax,aMax,0,0])
+ps = ProblemSolver( fullBody )
+ps.setParameter("Kinodynamic/velocityBound",vMax)
+ps.setParameter("Kinodynamic/accelerationBound",aMax)
+#load the viewer
+try :
+    v = Viewer (ps,viewerClient=v.client, displayCoM = True)
+except Exception:
+    print "No viewer started !"
+    class FakeViewer():
+        def __init__(self):
+            return
+        def __call__(self,q):
+            return
+        def addLandmark(self,a,b):
+            return
+    v = FakeViewer()
+
+
+if cfg.INIT_CONFIG_WB is None:
+    # load a reference configuration
+    q_ref = fullBody.referenceConfig[::]+[0]*6
+    #q_ref = fullBody.referenceConfig_legsApart[::]+[0]*6
+else:
+    q_ref = cfg.INIT_CONFIG_WB + [0]*6
+q_init = q_ref[::]
+fullBody.setReferenceConfig(q_ref)
+fullBody.setPostureWeights(fullBody.postureWeights[::]+[0]*6)
+fullBody.usePosturalTaskContactCreation(True)
+
+
+fullBody.setCurrentConfig (q_init)
+
+print "Generate limb DB ..."
+tStart = time.time()
+# generate databases : 
+"""
+nbSamples = 100000
+fullBody.addLimb(fullBody.rLegId,fullBody.rleg,fullBody.rfoot,fullBody.rLegOffset,fullBody.rLegNormal, fullBody.rLegx, fullBody.rLegy, nbSamples, heuristicR, 0.01,kinematicConstraintsPath=fullBody.rLegKinematicConstraints,kinematicConstraintsMin = 0.85)
+fullBody.runLimbSampleAnalysis(fullBody.rLegId, "ReferenceConfiguration", True)
+fullBody.addLimb(fullBody.lLegId,fullBody.lleg,fullBody.lfoot,fullBody.lLegOffset,fullBody.rLegNormal, fullBody.lLegx, fullBody.lLegy, nbSamples, heuristicL, 0.01,kinematicConstraintsPath=fullBody.lLegKinematicConstraints,kinematicConstraintsMin = 0.85)
+fullBody.runLimbSampleAnalysis(fullBody.lLegId, "ReferenceConfiguration", True)
+"""
+#~ fullBody.loadAllLimbs("fixedStep04","ReferenceConfiguration")
+fullBody.loadAllLimbs("static","ReferenceConfiguration")
+
+tGenerate =  time.time() - tStart
+print "Done."
+print "Databases generated in : "+str(tGenerate)+" s"
+
+#define initial and final configurations : 
+configSize = fullBody.getConfigSize() -fullBody.client.robot.getDimensionExtraConfigSpace()
+
+q_init[0:7] = cfg.INIT_CONFIG_ROOT[0:7] # use this to get the correct orientation
+vel_init = [0,0,0]
+acc_init = [0,0,0]
+vel_goal = [0,0,0]
+acc_goal = [0,0,0]
+
+robTreshold = 3
+# copy extraconfig for start and init configurations
+q_init[configSize:configSize+3] = vel_init[::]
+q_init[configSize+3:configSize+6] = acc_init[::]
+#~ q_goal[configSize:configSize+3] = vel_goal[::]
+#~ q_goal[configSize+3:configSize+6] = [0,0,0]
+
+
+q_init[2] = q_ref[2]
+#~ q_goal[2] = q_ref[2]
+
+
+
+
+fullBody.setStaticStability(True)
+fullBody.setCurrentConfig (q_init)
+v(q_init)
+
+
+v.addLandmark('anymal/base_0',0.3)
+v(q_init)
+#fullBody.setReferenceConfig(fullBody.referenceConfig_legsApart[::]+[0]*6)
+
+fullBody.setStartState(q_init,[fullBody.rArmId,fullBody.rLegId,fullBody.lArmId,fullBody.lLegId],[[0.,0.,1.] for _ in range(4)])
+
+from hpp.corbaserver.rbprm import rbprmstate
+from hpp.corbaserver.rbprm.rbprmstate import State
+
+initState = State(fullBody, fullBody.getNumStates()-1)
+
+
+import hpp.corbaserver.rbprm.fewstepsplanner as sp
+
+def dispContactPlan(states, step = 0.5):
+	for s in states:
+		v(s.q());
+		time.sleep(step)
+
+fewStepPlanner = sp.FewStepPlanner(cl,psGuide,rbprmBuilder, fullBody, pathPlayer = pp)

src/hpp/corbaserver/rbprm/fewstepsplanner.py

@@ -24,26 +24,38 @@ from numpy import array
 from hpp.corbaserver.rbprm import rbprmstate
 from hpp.corbaserver.rbprm.rbprmstate import State
 
-def interpolateState(fullBody, stepsize, pathId = 1, robustnessTreshold = 0, filterStates = False, testReachability = True, quasiStatic = False, erasePreviousStates = False):
+def _interpolateState(ps, fullBody, stepsize, pathId, robustnessTreshold = 0, filterStates = False, testReachability = True, quasiStatic = False, erasePreviousStates = True):
   if(filterStates):
         filt = 1
   else:
         filt = 0
-  configs = fullBody.clientRbprm.rbprm.interpolate(stepsize, pathId, robustnessTreshold, filt, testReachability, quasiStatic, erasePreviousStates)
+        
+  #discretize path
+  totalLength = ps.pathLength(pathId)
+  configsPlan = []; step = 0.
+  configSize = fullBody.getConfigSize() - len(ps.configAtParam(pathId,0.))
+  z = [0. for _ in range(configSize) ]
+  while step < totalLength:
+    configsPlan += [ps.configAtParam(pathId,step) + z[:]]
+    step += stepsize
+  configsPlan += [ps.configAtParam(pathId,totalLength)+ z[:]]
+    
+  configs = fullBody.clientRbprm.rbprm.interpolateConfigs(configsPlan, robustnessTreshold, filt, testReachability, quasiStatic, erasePreviousStates)
   firstStateId = fullBody.clientRbprm.rbprm.getNumStates() - len(configs)
-  return [ State(fullBody, i) for i in range(firstStateId, firstStateId + len(configs)) ]
+  return [ State(fullBody, i) for i in range(firstStateId, firstStateId + len(configs)) ], configs
           
 
-def guidePath(problemSolver, fromPos, toPos):
+def _guidePath(problemSolver, fromPos, toPos):
   ps = problemSolver
   ps.setInitialConfig (fromPos)
+  ps.resetGoalConfigs()
   ps.addGoalConfig(toPos)
   ps.solve ()
   return ps.numberPaths() - 1
   
 
 class FewStepPlanner(object):
-  def __init__ (self, client, problemSolver, rbprmBuilder, fullBody, planContext="rbprm_path", fullBodyContext="default", pathPlayer = None ):
+  def __init__ (self, client, problemSolver, rbprmBuilder, fullBody, planContext="rbprm_path", fullBodyContext="default", pathPlayer = None, viewer = None ):
     self.fullBody =  fullBody
     self.rbprmBuilder =  rbprmBuilder
     self.client   =  client
@@ -51,6 +63,7 @@ class FewStepPlanner(object):
     self.fullBodyContext   =  fullBodyContext
     self.problemSolver   =  problemSolver
     self.pathPlayer   =  pathPlayer
+    self.viewer   =  viewer
     
   def setPlanningContext(self):
     self.client.problem.selectProblem(self.planContext) 
@@ -71,13 +84,32 @@ class FewStepPlanner(object):
     self.setCurrentContext(oldContext)
     return res
     
-  def guidePath(self, fromPos, toPos):
-    pId =  self._actInContext(self.planContext,guidePath,self.problemSolver, fromPos, toPos)
+  def guidePath(self, fromPos, toPos, displayPath = False):
+    pId =  self._actInContext(self.planContext,_guidePath,self.problemSolver, fromPos, toPos)
     self.setPlanningContext()
     names =  self.rbprmBuilder.getAllJointNames()[1:]
-    self.pathPlayer(pId)
+    if displayPath:
+      if self.pathPlayer is None:
+        print "can't display path, no path player given"
+      else:
+        self.pathPlayer(pId)        
     self.client.problem.movePathToProblem(pId,self.fullBodyContext, names)
     self.setFullBodyContext()
     return pId
+    
+  def interpolateStates(self, stepsize, pathId = 1, robustnessTreshold = 0, filterStates = False, testReachability = True, quasiStatic = False, erasePreviousStates = True):
+    return _interpolateState(self.problemSolver, self.fullBody, stepsize, pathId, robustnessTreshold, filterStates, testReachability, quasiStatic, erasePreviousStates)
+    
+  def goToQuasiStatic(self, currentState, toPos, stepsize = 0.002, goalLimbsInContact = None, goalNormals = None, displayGuidePath = False):
+    pId = self.guidePath(currentState.q()[:7], toPos, displayPath = displayGuidePath)    
+    self.fullBody.setStartStateId(currentState.sId)
+    targetState = currentState.q()[:]; targetState[:7] = toPos
+    if goalLimbsInContact is None:
+      goalLimbsInContact = currentState.getLimbsInContact()
+    if goalNormals is None:
+      goalNormals = [[0.,0.,1.] for _ in range(len(goalLimbsInContact))]
+    self.fullBody.setEndState(targetState, goalLimbsInContact,goalNormals)
+    return self.interpolateStates(stepsize,pathId=pId,robustnessTreshold = 1, filterStates = True,quasiStatic=True, erasePreviousStates = False)
+    
        
               

src/rbprmbuilder.impl.cc

@@ -1574,8 +1574,8 @@ namespace hpp {
             std::size_t id = 0;
             for(std::vector<State>::const_iterator cit = newStates.begin(); cit != newStates.end(); ++cit, ++id)
             {
-                std::cout << "ID " << id;
-                cit->print();
+                /*std::cout << "ID " << id;
+                cit->print();*/
                 const core::Configuration_t config = cit->configuration_;
                 _CORBA_ULong size = (_CORBA_ULong) config.size ();
                 double* dofArray = hpp::floatSeq::allocbuf(size);