forked from agimus/agimus-demos-franka-manipulation
-
Notifications
You must be signed in to change notification settings - Fork 0
/
bin_picking.py
550 lines (513 loc) · 22.7 KB
/
bin_picking.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
# Copyright 2023 CNRS
# Author: Florent Lamiraux
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
# 1. Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# 2. Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
from hpp.corbaserver import Client, loadServerPlugin, shrinkJointRange, \
wrap_delete
from hpp.corbaserver.bin_picking import Client as BpClient
from hpp.corbaserver.manipulation import Client as ManipClient, ProblemSolver
from hpp.corbaserver.manipulation import Constraints, Robot, Rule
from hpp.corbaserver.problem_solver import _convertToCorbaAny as convertToAny
from agimus_demos import InStatePlanner
from create_graph import makeGraph
from agimus_demos.tools_hpp import concatenatePaths
from tools_hpp import displayGripper, displayHandle, generateTargetConfig, \
shootPartInBox
## Write a handle in a string in srdf format
def writeHandleInSrdf(robot, handle, clearance, mask):
joint, pose = robot.getHandlePositionInJoint(handle)
link = robot.getLinkNames(joint)[0]
# Guess prefix
rank = handle.find("/")
if rank == -1:
prefix = ""
handle_short = handle
link_short = link
else:
prefix = handle[:rank]
handle_short = handle[rank+1:]
if link[:rank] != prefix:
raise RuntimeError(
f"Expected prefix {prefix}/ in link name, but got {link}")
link_short = link[rank+1:]
res = f'''
<handle name="{handle_short}" clearance="{clearance}">
<position xyz="{pose[0]} {pose[1]} {pose[2]}" wxyz="{pose[6]} {pose[3]} {pose[4]} {pose[5]}"/>
<link name="{link_short}"/>
<mask>{mask[0]} {mask[1]} {mask[2]} {mask[3]} {mask[4]} {mask[5]}</mask>
</handle>
'''
return res
## This class defines a bin-picking problem.
#
class BinPicking(object):
objects = list()
"""
List of object names. The part to grasp is the first one.
"""
robotGrippers = list()
"""
List of robot grippers to grasp the object
"""
goalGrippers = list()
"""
List of gripper that define goal poses of the object, ranked in order of
priority. Using only one robot gripper, it may not be possible to release
the object always in the same pose.
"""
nh = 16
"""
Number of handles to produce when discretizing
"""
handles = list()
"""
List of handles of the object to grasp
"""
handlesToDiscretize = list()
"""
List of handles to discretize
"""
goalHandles = list()
"""
List of handles that define the goal poses of the object. mapped one to
one with goalGrippers.
"""
graphConstraints = list()
"""
List of constraints to be added to the constraint graph: quasi-static
equilibrium, gripper open...
"""
goalConfigs = dict()
"""
List of precomputed goal configurations to release the object
"""
q_goal = None
"""
Configuration with which goal configurations are precomputed.
This configuration defines the poses of objects other than the part.
"""
timeParamDict = {'freefly':{'order' : 2, 'maxAcc' : 1., 'safety' : 0.95},'grasping':{'order' : 2, 'maxAcc' : 0.1, 'safety' : 0.95},'approach':{'order' : 2, 'maxAcc' : 0.5, 'safety' : 0.95}}
"""
Configuration of end effector for freefly or grasp.
"""
def wd(self, o):
return wrap_delete(o, self.ps.client.basic._tools)
def __init__(self, ps):
self.ps = ps
self.robot = ps.robot
self.bpc = BpClient()
# Store place paths where the object is released
self.placePaths = dict()
def _rules(self):
"""
Build rules for the constraint graph factory
"""
rules = list()
for ig, gripper in enumerate(self.goalGrippers):
for ih, handle in enumerate(self.goalHandles):
if ig != ih:
rules += [
Rule(grippers=[gripper], handles = [handle], link=False)
]
# object cannot be placed in two goal positions at the same time
rules += [
Rule(grippers=self.robotGrippers + self.goalGrippers,
handles = [".*", "part/center1", "part/center2"],
link=False),
Rule(grippers=[".*"], handles=[".*"], link=True)
]
return rules
def _possibleGrasps(self):
res = dict()
# Each goal gripper can only grasp one goal handle
if len(self.goalGrippers) != len(self.goalHandles):
raise RuntimeError(
f"number of goal grippers ({len(self.goalGrippers)})"
+ " should be the same as " +
f"number of goal handles ({len(self.goalHandles)})"
)
for g,h in zip(self.goalGrippers, self.goalHandles):
res[g] = [h]
for g in self.robotGrippers:
res[g] = self.handles
return res
def writeRules(self, f):
"""
Write the rules in a stream
"""
f.write("rules:\n")
for r in self._rules():
f.write(f" - grippers: {r.grippers}\n" +
f" handles: {r.handles}\n" +
f" link: {r.link}\n")
def writeTranstions(self, f):
"""
Write the edges in a yaml file
"""
f.write("transitions:\n")
for e in self.graph.edges.keys():
f.write(f' - "{e}"\n')
def writeStates(self, f):
"""
Write the nodes of the constraint graph in a yaml file
"""
f.write("states:\n")
for s in self.graph.nodes.keys():
f.write(f' - "{s}"\n')
def buildGraph(self):
"""
Build the constraint graph
- discretize handles,
"""
# store list of not discretized handles
self.initialHandles = self.handles[:]
# Discretize handles
for h in self.handlesToDiscretize:
self.bpc.bin_picking.discretizeHandle(h, self.nh)
self.handles += ['%s_%03d' % (h, i) for i in range(self.nh)]
# build graph
# Sort handles in alphabetic order to be sure that the order is the
# same in agimus-sot.
handles = self.handles + self.goalHandles
handles.sort()
self.factory, self.graph = makeGraph(
self.ps, self.robot, self.robotGrippers + self.goalGrippers,
self.objects, [handles, []], self._rules(), self._possibleGrasps())
self.graph.addConstraints(graph=True,
constraints = Constraints(numConstraints = self.graphConstraints))
self.graph.initialize()
self.transitionPlanner = self.wd(self.ps.client.manipulation.problem.\
createTransitionPlanner())
self.transitionPlanner.timeOut(3.)
self.transitionPlanner.setPathProjector('Progressive', 0.2)
self.transitionPlanner.addPathOptimizer("EnforceTransitionSemantic")
self.transitionPlanner.addPathOptimizer("SimpleTimeParameterization")
def buildEffectors(self, obstacles, q):
"""
build and effector to test collision of grasps
- name: name of the effector,
- gripper: name of the gripper,
- obstacles: list of obstacle name with which the effector with
be tested for collision,
- q: any configuration where the various geometries of the gripper are
in the right relative pose.
"""
for gripper in self.robotGrippers:
# find any edge that grasps the object
edge = None
for e in self.graph.edges.keys():
if e.startswith(gripper + " > ") and e.endswith("12"):
edge = e
break
if edge is None:
raise RuntimeError(
f"Did not find any edge where {gripper} grasps anything")
self.bpc.bin_picking.createEffector(gripper, gripper, q,
self.graph.edges[edge])
for o in obstacles:
self.bpc.bin_picking.addObstacleToEffector(gripper, o, 0.)
def generateGoalConfigs(self, q):
"""
Precompute goal configurations for the part hold by the robot
- q configuration that provides the pose of other objects
"""
self.q_goal = q[:]
# Separate indices of handles and grippers to distinguish goal
# grippers/handles from regular ones
robotGrippers = list()
goalGrippers = list()
regularHandles = list()
goalHandles = list()
for i, h in enumerate(self.factory.handles):
if h in self.handles:
regularHandles.append(i)
elif h in self.goalHandles:
goalHandles.append(i)
else:
raise RuntimeError(
f"'{h}' is neither a regular handle nor a goal handle")
for i, g in enumerate(self.factory.grippers):
if g in self.robotGrippers:
robotGrippers.append(i)
elif g in self.goalGrippers:
goalGrippers.append(i)
else:
raise RuntimeError(
f"'{g}' is neither a regular gripper nor a goal gripper")
for irg in robotGrippers:
robotGripper = self.factory.grippers[irg]
self.placePaths[robotGripper] = dict()
for ih in regularHandles:
handle = self.factory.handles[ih]
found = False
for igg, (goalGripper, goalHandle) in enumerate(
zip(self.goalGrippers, self.goalHandles)):
ggIndex = self.factory.grippers.index(goalGripper)
ghIndex = self.factory.handles.index(goalHandle)
edges = [f"{goalGripper} > {goalHandle} | {irg}-{ih}_01",
f"{goalGripper} > {goalHandle} | {irg}-{ih}_12",
f"{robotGripper} < {handle} | {irg}-{ih}:" + \
f"{ggIndex}-{ghIndex}_21"]
p = self.generateConsecutivePaths(edges, q)
if p:
self.placePaths[robotGripper][handle] = p
found = True
break
if found: break
def generateConsecutivePaths(self, edges, q, Nsamples = 50):
"""
Generate consecutive paths along a list of edges
The paths starts along the second edge. The first edge is used
to generate the initial configuration of the path.
The constraint right hand sides are initialized with
- the input configuration q for the first edge,
- the end of the previous path for the following edges
"""
for i in range(Nsamples + 1):
if i == 0:
qrand = q[:]
else:
qrand = self.robot.shootRandomConfig()
r = self.robot.rankInConfiguration['box/root_joint']
qrand[r:r+7] = q[r:r+7]
waypoints = list()
success = True
for edge in edges:
q1 = generateTargetConfig(self.robot, self.graph, edge, q, qrand)
if not q1:
success = False
break
waypoints.append(q1)
q = q1[:]; qrand = q1[:]
# If the waypoints have been successfully generated, we
# plan paths between them
paths = list()
if success:
for q1, q2, edge in zip(waypoints, waypoints[1:], edges[1:]):
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('grasping')
p, res, msg = self.transitionPlanner.directPath(q1, q2,
True)
if not res:
success = False
break
else:
p = self.wd(p)
p = self.transitionPlanner.timeParameterization(p.asVector())
paths.append(self.wd(p))
if success:
return concatenatePaths(paths)
return None
def checkObjectPoses(self, q):
"""
Check that
- the initial configuration is in state "free",
- each object is in the same pose as in the configuration
that was used to precompute the goal configurations. Index 0
corresponds to the part and thus is skipped.
"""
if self.graph.getNode(q) != "free":
raise RuntimeError(
f"Initial configuration {q} not in state 'free'.")
if not self.q_goal:
raise RuntimeError(
"You need to call method generateGoalConfigs first.")
for o in self.objects[1:]:
r = self.robot.rankInConfiguration[f"{o}/root_joint"]
if q[r:r+7] != self.q_goal[r:r+7]:
raise RuntimeError(f"Object {o} is in pose {q[r:r+7]} but " +
"was in pose {self.q_goal[r:r+7]} when pre-computing" +
" goal configurations.")
def computeFreeGrasps(self, q):
"""
For each gripper, compute list of object handles that are not in
collision in the given configuration.
Result is a dictionary of keys the robot grippers and with values
lists of handles.
"""
self._freeGrasps = dict()
res = False
for gripper in self.robotGrippers:
freeGrasps = list()
self._freeGrasps[gripper] = list()
for handle in self.handles:
col, msg, gripperAxis = self.bpc.bin_picking.collisionTest(
gripper, handle, q)
if not col:
# Store pairs (handle, score)
freeGrasps.append((handle, -gripperAxis[2]))
res = True
# Sort handles by increasing z coordinate of gripper axis
l = sorted(freeGrasps, key = lambda x:x[1], reverse = True)
if len(l) > 0:
self._freeGrasps[gripper] = list(zip(*l))[0]
return res
def selectGrasp(self, q):
"""
Select a grasp among all possible collision-free grasps
- q initial configuration of the system,
- return gripper, handle, and 6 waypoint configurations corresponding
to - pregrasp, grasp, preplace configurations to grasp the object,
- pregrasp, grasp, preplace configurations to release the
object.
"""
for gripper in self.robotGrippers:
for handle in self._freeGrasps[gripper]:
# check that place path exists for this grasp
placePath = self.placePaths[gripper].get(handle)
if not placePath:
break
# generate pregrasp, grasp and preplace configuration to
# grasp the object
edges = [f"{gripper} > {handle} | f_01",
f"{gripper} > {handle} | f_12",
f"{gripper} > {handle} | f_23"]
pickPath = self.generateConsecutivePaths(edges, q)
if pickPath:
return gripper, handle, pickPath, placePath
return 4*(None,)
def setParam(self, state):
if state not in self.timeParamDict:
raise RuntimeError(f"state value should belong to {list(self.timeParamDict.keys())}")
self.transitionPlanner.setParameter('SimpleTimeParameterization/order', convertToAny(self.timeParamDict[state]['order']))
self.transitionPlanner.setParameter('SimpleTimeParameterization/maxAcceleration', convertToAny(self.timeParamDict[state]['maxAcc']))
self.transitionPlanner.setParameter('SimpleTimeParameterization/safety', convertToAny(self.timeParamDict[state]['safety']))
def direct_Path(self, q_start, q_goal):
q = q_start
edge = "Loop | f"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('approach')
try:
p_direct, res, msg = self.transitionPlanner.directPath(q_start, q_goal, True)
print("Achieve to create a direct path : ",res)
if not res:
gripper, handle, pickPath, placePath = self.selectGrasp(q)
q1 = pickPath.initial()
p_direct = self.wd(self.transitionPlanner.planPath(q, [q1,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q_start} and {q_goal}: {exc}")
return True, p_direct
def solve(self, q, type_of_path='default'):
"""
Compute a trajectory to grasp and release the object
- q initial configuration of the robot and objects
"""
self.checkObjectPoses(q)
res = self.computeFreeGrasps(q)
if not res:
msg = "End effector in collision for all possible grasps."
return False, msg
gripper, handle, pickPath, placePath = self.selectGrasp(q)
if gripper is None:
msg = "Failed to generate collision-free grasp with valid goal " +\
"configurations"
return False, msg
# Direct Path Plan
if type_of_path == 'direct_path':
edge = "Loop | f"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('approach')
try:
q1 = pickPath.initial()
p_direct = self.wd(self.transitionPlanner.planPath(q, [q1,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q} and {q1}: {exc}")
return True, p_direct
# Precise Path Plan
if type_of_path == 'precise':
ig = self.factory.grippers.index(gripper)
ih = self.factory.handles.index(handle)
edge = f"Loop | {ig}-{ih}"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('grasping')
try:
q3 = pickPath.end()
q4 = placePath.initial()
p4 = self.wd(self.transitionPlanner.planPath(q3, [q4,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q3} and {q4}: {exc}")
# Return to initial configuration
edge = "Loop | f"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('freefly')
q6 = placePath.end()
q7 = q[:]
r = self.robot.rankInConfiguration[f"{self.objects[0]}/root_joint"]
q7[r:r+7] = q6[r:r+7]
try:
p7 = self.wd(self.transitionPlanner.planPath(q6, [q7,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q6} and {q7}: {exc}")
return True, concatenatePaths([pickPath,p4,placePath,p7])
# Default Path Plan
if type_of_path == 'default':
# Plan paths between waypoint configurations
edge = "Loop | f"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('freefly')
try:
q1 = pickPath.initial()
p1 = self.wd(self.transitionPlanner.planPath(q, [q1,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q} and {q1}: {exc}")
ig = self.factory.grippers.index(gripper)
ih = self.factory.handles.index(handle)
edge = f"Loop | {ig}-{ih}"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('grasping')
try:
q3 = pickPath.end()
q4 = placePath.initial()
p4 = self.wd(self.transitionPlanner.planPath(q3, [q4,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q3} and {q4}: {exc}")
# Return to initial configuration
edge = "Loop | f"
self.transitionPlanner.setEdge(self.graph.edges[edge])
self.setParam('freefly')
q6 = placePath.end()
q7 = q[:]
r = self.robot.rankInConfiguration[f"{self.objects[0]}/root_joint"]
q7[r:r+7] = q6[r:r+7]
try:
p7 = self.wd(self.transitionPlanner.planPath(q6, [q7,], True))
except Exception as exc:
raise RuntimeError(f"Failed to connect {q6} and {q7}: {exc}")
return True, concatenatePaths([p1,pickPath,p4,placePath,p7])
if __name__ == "__main__":
from hpp.rostools import process_xacro, retrieve_resource
Robot.urdfString = process_xacro\
("package://agimus_demos/franka/manipulation/urdf/demo.urdf.xacro")
Robot.srdfString = ""
problems = ["default", "learning"]
defaultContext = "corbaserver"
loadServerPlugin (defaultContext, "manipulation-corba.so")
cl = ManipClient()
for p in problems:
res = cl.problem.selectProblem(p)
if res: print(f'created problem "{p}"')
else: print(f'did not create problem "{p}"')
cl.problem.resetProblem()
cl.problem.selectProblem("default")
robot = Robot("robot", "pandas", rootJointType="anchor")
shrinkJointRange(robot, [f'pandas/panda2_joint{i}' for i in range(1,8)],
0.95)
ps = ProblemSolver(robot)