Loading...
Searching...
No Matches
PRM.cpp
1/*********************************************************************
2 * Software License Agreement (BSD License)
3 *
4 * Copyright (c) 2011, Rice University
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 *
11 * * Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * * Redistributions in binary form must reproduce the above
14 * copyright notice, this list of conditions and the following
15 * disclaimer in the documentation and/or other materials provided
16 * with the distribution.
17 * * Neither the name of Rice University nor the names of its
18 * contributors may be used to endorse or promote products derived
19 * from this software without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
25 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
27 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
29 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
31 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
32 * POSSIBILITY OF SUCH DAMAGE.
33 *********************************************************************/
34
35/* Author: Ioan Sucan, James D. Marble, Ryan Luna, Henning Kayser */
36
37#include "ompl/geometric/planners/prm/PRM.h"
38#include "ompl/geometric/planners/prm/ConnectionStrategy.h"
39#include "ompl/base/goals/GoalSampleableRegion.h"
40#include "ompl/base/objectives/PathLengthOptimizationObjective.h"
41#include "ompl/datastructures/PDF.h"
42#include "ompl/tools/config/SelfConfig.h"
43#include "ompl/tools/config/MagicConstants.h"
44#include <boost/graph/astar_search.hpp>
45#include <boost/graph/incremental_components.hpp>
46#include <boost/property_map/vector_property_map.hpp>
47#include <boost/foreach.hpp>
48#include <thread>
49#include <typeinfo>
50
51#include "GoalVisitor.hpp"
52
53#define foreach BOOST_FOREACH
54
55namespace ompl
56{
57 namespace magic
58 {
61 static const unsigned int MAX_RANDOM_BOUNCE_STEPS = 5;
62
64 static const double ROADMAP_BUILD_TIME = 0.2;
65
68 static const unsigned int DEFAULT_NEAREST_NEIGHBORS = 10;
69 } // namespace magic
70} // namespace ompl
71
72ompl::geometric::PRM::PRM(const base::SpaceInformationPtr &si, bool starStrategy)
73 : base::Planner(si, "PRM")
74 , starStrategy_(starStrategy)
75 , stateProperty_(boost::get(vertex_state_t(), g_))
78 , weightProperty_(boost::get(boost::edge_weight, g_))
79 , disjointSets_(boost::get(boost::vertex_rank, g_), boost::get(boost::vertex_predecessor, g_))
80{
81 specs_.recognizedGoal = base::GOAL_SAMPLEABLE_REGION;
82 specs_.approximateSolutions = true;
83 specs_.optimizingPaths = true;
84 specs_.multithreaded = true;
85
86 if (!starStrategy_)
87 Planner::declareParam<unsigned int>("max_nearest_neighbors", this, &PRM::setMaxNearestNeighbors,
88 &PRM::getMaxNearestNeighbors, std::string("8:1000"));
89
90 addPlannerProgressProperty("iterations INTEGER", [this] { return getIterationCount(); });
91 addPlannerProgressProperty("best cost REAL", [this] { return getBestCost(); });
92 addPlannerProgressProperty("milestone count INTEGER", [this] { return getMilestoneCountString(); });
93 addPlannerProgressProperty("edge count INTEGER", [this] { return getEdgeCountString(); });
94}
95
96ompl::geometric::PRM::PRM(const base::PlannerData &data, bool starStrategy)
97 : PRM(data.getSpaceInformation(), starStrategy)
98{
99 if (data.numVertices() > 0)
100 {
101 // mapping between vertex id from PlannerData and Vertex in Boost.Graph
102 std::map<unsigned int, Vertex> vertices;
103 // helper function to create vertices as needed and update the vertices mapping
104 const auto &getOrCreateVertex = [&](unsigned int vertex_index) {
105 if (!vertices.count(vertex_index))
106 {
107 const auto &data_vertex = data.getVertex(vertex_index);
108 Vertex graph_vertex = boost::add_vertex(g_);
109 stateProperty_[graph_vertex] = si_->cloneState(data_vertex.getState());
110 totalConnectionAttemptsProperty_[graph_vertex] = 1;
111 successfulConnectionAttemptsProperty_[graph_vertex] = 0;
112 vertices[vertex_index] = graph_vertex;
113 }
114 return vertices.at(vertex_index);
115 };
116
117 specs_.multithreaded = false; // temporarily set to false since nn_ is used only in single thread
119 specs_.multithreaded = true;
120 nn_->setDistanceFunction([this](const Vertex a, const Vertex b) { return distanceFunction(a, b); });
121
122 for (size_t vertex_index = 0; vertex_index < data.numVertices(); ++vertex_index)
123 {
124 Vertex m = getOrCreateVertex(vertex_index);
125 std::vector<unsigned int> neighbor_indices;
126 data.getEdges(vertex_index, neighbor_indices);
127 if (neighbor_indices.empty())
128 {
129 disjointSets_.make_set(m);
130 }
131 else
132 {
133 for (const unsigned int neighbor_index : neighbor_indices)
134 {
135 Vertex n = getOrCreateVertex(neighbor_index);
138 base::Cost weight;
139 data.getEdgeWeight(vertex_index, neighbor_index, &weight);
140 const Graph::edge_property_type properties(weight);
141 boost::add_edge(m, n, properties, g_);
142 uniteComponents(m, n);
143 }
144 }
145 nn_->add(m);
146 }
147 }
148}
149
150ompl::geometric::PRM::~PRM()
151{
152 freeMemory();
153}
154
156{
157 Planner::setup();
158 if (!nn_)
159 {
160 specs_.multithreaded = false; // temporarily set to false since nn_ is used only in single thread
162 specs_.multithreaded = true;
163 nn_->setDistanceFunction([this](const Vertex a, const Vertex b) { return distanceFunction(a, b); });
164 }
168 connectionFilter_ = [](const Vertex &, const Vertex &) { return true; };
169
170 // Setup optimization objective
171 //
172 // If no optimization objective was specified, then default to
173 // optimizing path length as computed by the distance() function
174 // in the state space.
175 if (pdef_)
176 {
177 if (pdef_->hasOptimizationObjective())
178 opt_ = pdef_->getOptimizationObjective();
179 else
180 {
181 opt_ = std::make_shared<base::PathLengthOptimizationObjective>(si_);
182 if (!starStrategy_)
183 opt_->setCostThreshold(opt_->infiniteCost());
184 }
185 }
186 else
187 {
188 OMPL_INFORM("%s: problem definition is not set, deferring setup completion...", getName().c_str());
189 setup_ = false;
190 }
191}
192
194{
195 if (starStrategy_)
196 throw Exception("Cannot set the maximum nearest neighbors for " + getName());
197 if (!nn_)
198 {
199 specs_.multithreaded = false; // temporarily set to false since nn_ is used only in single thread
201 specs_.multithreaded = true;
202 nn_->setDistanceFunction([this](const Vertex a, const Vertex b) { return distanceFunction(a, b); });
203 }
206 if (isSetup())
207 setup();
208}
209
211{
212 const auto strategy = connectionStrategy_.target<KStrategy<Vertex>>();
213 return strategy ? strategy->getNumNeighbors() : 0u;
214}
215
223
224void ompl::geometric::PRM::setProblemDefinition(const base::ProblemDefinitionPtr &pdef)
225{
226 Planner::setProblemDefinition(pdef);
227 clearQuery();
228}
229
231{
232 startM_.clear();
233 goalM_.clear();
234 pis_.restart();
235}
236
238{
239 Planner::clear();
240 sampler_.reset();
241 simpleSampler_.reset();
242 freeMemory();
243 if (nn_)
244 nn_->clear();
245 clearQuery();
246
247 iterations_ = 0;
248 bestCost_ = base::Cost(std::numeric_limits<double>::quiet_NaN());
249}
250
252{
253 foreach (Vertex v, boost::vertices(g_))
254 si_->freeState(stateProperty_[v]);
255 g_.clear();
256}
257
262
264{
265 if (!simpleSampler_)
266 simpleSampler_ = si_->allocStateSampler();
267
268 std::vector<base::State *> states(magic::MAX_RANDOM_BOUNCE_STEPS);
269 si_->allocStates(states);
270 expandRoadmap(ptc, states);
271 si_->freeStates(states);
272}
273
275 std::vector<base::State *> &workStates)
276{
277 // construct a probability distribution over the vertices in the roadmap
278 // as indicated in
279 // "Probabilistic Roadmaps for Path Planning in High-Dimensional Configuration Spaces"
280 // Lydia E. Kavraki, Petr Svestka, Jean-Claude Latombe, and Mark H. Overmars
281
282 PDF<Vertex> pdf;
283 foreach (Vertex v, boost::vertices(g_))
284 {
285 const unsigned long int t = totalConnectionAttemptsProperty_[v];
286 pdf.add(v, (double)(t - successfulConnectionAttemptsProperty_[v]) / (double)t);
287 }
288
289 if (pdf.empty())
290 return;
291
292 while (!ptc)
293 {
294 iterations_++;
295 Vertex v = pdf.sample(rng_.uniform01());
296 unsigned int s =
297 si_->randomBounceMotion(simpleSampler_, stateProperty_[v], workStates.size(), workStates, false);
298 if (s > 0)
299 {
300 s--;
301 Vertex last = addMilestone(si_->cloneState(workStates[s]));
302
303 graphMutex_.lock();
304 for (unsigned int i = 0; i < s; ++i)
305 {
306 // add the vertex along the bouncing motion
307 Vertex m = boost::add_vertex(g_);
308 stateProperty_[m] = si_->cloneState(workStates[i]);
311 disjointSets_.make_set(m);
312
313 // add the edge to the parent vertex
314 const base::Cost weight = opt_->motionCost(stateProperty_[v], stateProperty_[m]);
315 const Graph::edge_property_type properties(weight);
316 boost::add_edge(v, m, properties, g_);
317 uniteComponents(v, m);
318
319 // add the vertex to the nearest neighbors data structure
320 nn_->add(m);
321 v = m;
322 }
323
324 // if there are intermediary states or the milestone has not been connected to the initially sampled vertex,
325 // we add an edge
326 if (s > 0 || !sameComponent(v, last))
327 {
328 // add the edge to the parent vertex
329 const base::Cost weight = opt_->motionCost(stateProperty_[v], stateProperty_[last]);
330 const Graph::edge_property_type properties(weight);
331 boost::add_edge(v, last, properties, g_);
332 uniteComponents(v, last);
333 }
334 graphMutex_.unlock();
335 }
336 }
337}
338
343
345{
346 if (!isSetup())
347 setup();
348 if (!sampler_)
349 sampler_ = si_->allocValidStateSampler();
350
351 base::State *workState = si_->allocState();
352 growRoadmap(ptc, workState);
353 si_->freeState(workState);
354}
355
357{
358 /* grow roadmap in the regular fashion -- sample valid states, add them to the roadmap, add valid connections */
359 while (!ptc)
360 {
361 iterations_++;
362 // search for a valid state
363 bool found = false;
364 while (!found && !ptc)
365 {
366 unsigned int attempts = 0;
367 do
368 {
369 found = sampler_->sample(workState);
370 attempts++;
372 }
373 // add it as a milestone
374 if (found)
375 addMilestone(si_->cloneState(workState));
376 }
377}
378
380{
381 auto *goal = static_cast<base::GoalSampleableRegion *>(pdef_->getGoal().get());
382 while (!ptc && !addedNewSolution_)
383 {
384 // Check for any new goal states
385 if (goal->maxSampleCount() > goalM_.size())
386 {
387 const base::State *st = pis_.nextGoal();
388 if (st != nullptr)
389 goalM_.push_back(addMilestone(si_->cloneState(st)));
390 }
391
392 // Check for a solution
394 // Sleep for 1ms
396 std::this_thread::sleep_for(std::chrono::milliseconds(1));
397 }
398}
399
400bool ompl::geometric::PRM::maybeConstructSolution(const std::vector<Vertex> &starts, const std::vector<Vertex> &goals,
401 base::PathPtr &solution)
402{
403 base::Goal *g = pdef_->getGoal().get();
404 base::Cost sol_cost(opt_->infiniteCost());
405 foreach (Vertex start, starts)
406 {
407 foreach (Vertex goal, goals)
408 {
409 // we lock because the connected components algorithm is incremental and may change disjointSets_
410 graphMutex_.lock();
411 bool same_component = sameComponent(start, goal);
412 graphMutex_.unlock();
413
414 if (same_component && g->isStartGoalPairValid(stateProperty_[goal], stateProperty_[start]))
415 {
416 base::PathPtr p = constructSolution(start, goal);
417 if (p)
418 {
419 base::Cost pathCost = p->cost(opt_);
420 if (opt_->isCostBetterThan(pathCost, bestCost_))
421 bestCost_ = pathCost;
422 // Check if optimization objective is satisfied
423 if (opt_->isSatisfied(pathCost))
424 {
425 solution = p;
426 return true;
427 }
428 if (opt_->isCostBetterThan(pathCost, sol_cost))
429 {
430 solution = p;
431 sol_cost = pathCost;
432 }
433 }
434 }
435 }
436 }
437
438 return false;
439}
440
445
447{
449 auto *goal = dynamic_cast<base::GoalSampleableRegion *>(pdef_->getGoal().get());
450
451 if (goal == nullptr)
452 {
453 OMPL_ERROR("%s: Unknown type of goal", getName().c_str());
455 }
456
457 // Add the valid start states as milestones
458 while (const base::State *st = pis_.nextStart())
459 startM_.push_back(addMilestone(si_->cloneState(st)));
460
461 if (startM_.empty())
462 {
463 OMPL_ERROR("%s: There are no valid initial states!", getName().c_str());
465 }
466
467 if (!goal->couldSample())
468 {
469 OMPL_ERROR("%s: Insufficient states in sampleable goal region", getName().c_str());
471 }
472
473 // Ensure there is at least one valid goal state
474 if (goal->maxSampleCount() > goalM_.size() || goalM_.empty())
475 {
476 const base::State *st = goalM_.empty() ? pis_.nextGoal(ptc) : pis_.nextGoal();
477 if (st != nullptr)
478 goalM_.push_back(addMilestone(si_->cloneState(st)));
479
480 if (goalM_.empty())
481 {
482 OMPL_ERROR("%s: Unable to find any valid goal states", getName().c_str());
484 }
485 }
486
487 unsigned long int nrStartStates = boost::num_vertices(g_);
488 OMPL_INFORM("%s: Starting planning with %lu states already in datastructure", getName().c_str(), nrStartStates);
489
490 // Reset addedNewSolution_ member and create solution checking thread
491 addedNewSolution_ = false;
492 base::PathPtr sol;
493 std::thread slnThread([this, &ptc, &sol] { checkForSolution(ptc, sol); });
494
495 // construct new planner termination condition that fires when the given ptc is true, or a solution is found
496 base::PlannerTerminationCondition ptcOrSolutionFound([this, &ptc] { return ptc || addedNewSolution(); });
497
498 constructRoadmap(ptcOrSolutionFound);
499
500 // Ensure slnThread is ceased before exiting solve
501 slnThread.join();
502
503 OMPL_INFORM("%s: Created %u states", getName().c_str(), boost::num_vertices(g_) - nrStartStates);
504
505 if (sol)
506 {
507 base::PlannerSolution psol(sol);
508 psol.setPlannerName(getName());
509 // if the solution was optimized, we mark it as such
511 pdef_->addSolutionPath(psol);
512 }
513 else
514 {
515 // Return an approximate solution.
517 if (!opt_->isFinite(diff))
518 {
519 OMPL_INFORM("Closest path is still start and goal");
521 }
522 OMPL_INFORM("Using approximate solution, heuristic cost-to-go is %f", diff.value());
523 pdef_->addSolutionPath(sol, true, diff.value(), getName());
525 }
526
528}
529
531{
532 if (!isSetup())
533 setup();
534 if (!sampler_)
535 sampler_ = si_->allocValidStateSampler();
536 if (!simpleSampler_)
537 simpleSampler_ = si_->allocStateSampler();
538
539 std::vector<base::State *> xstates(magic::MAX_RANDOM_BOUNCE_STEPS);
540 si_->allocStates(xstates);
541 bool grow = true;
542
543 bestCost_ = opt_->infiniteCost();
544 while (!ptc())
545 {
546 // maintain a 2:1 ratio for growing/expansion of roadmap
547 // call growRoadmap() twice as long for every call of expandRoadmap()
548 if (grow)
551 xstates[0]);
552 else
555 xstates);
556 grow = !grow;
557 }
558
559 si_->freeStates(xstates);
560}
561
563{
564 std::lock_guard<std::mutex> _(graphMutex_);
565
566 Vertex m = boost::add_vertex(g_);
567 stateProperty_[m] = state;
570
571 // Initialize to its own (dis)connected component.
572 disjointSets_.make_set(m);
573
574 // Which milestones will we attempt to connect to?
575 const std::vector<Vertex> &neighbors = connectionStrategy_(m);
576
577 foreach (Vertex n, neighbors)
578 if (connectionFilter_(n, m))
579 {
582 if (si_->checkMotion(stateProperty_[n], stateProperty_[m]))
583 {
586 const base::Cost weight = opt_->motionCost(stateProperty_[n], stateProperty_[m]);
587 const Graph::edge_property_type properties(weight);
588 boost::add_edge(n, m, properties, g_);
589 uniteComponents(n, m);
590 }
591 }
592
593 nn_->add(m);
594
595 return m;
596}
597
599{
600 disjointSets_.union_set(m1, m2);
601}
602
604{
605 return boost::same_component(m1, m2, disjointSets_);
606}
607
609 const std::vector<Vertex> &goals,
610 base::PathPtr &solution)
611{
612 std::lock_guard<std::mutex> _(graphMutex_);
613 base::Goal *g = pdef_->getGoal().get();
614 base::Cost closestVal(opt_->infiniteCost());
615 bool approxPathJustStart = true;
616
617 foreach (Vertex start, starts)
618 {
619 foreach (Vertex goal, goals)
620 {
621 base::Cost heuristicCost(costHeuristic(start, goal));
622 if (opt_->isCostBetterThan(heuristicCost, closestVal))
623 {
624 closestVal = heuristicCost;
625 approxPathJustStart = true;
626 }
628 {
629 continue;
630 }
631 base::PathPtr p;
632 boost::vector_property_map<Vertex> prev(boost::num_vertices(g_));
633 boost::vector_property_map<base::Cost> dist(boost::num_vertices(g_));
634 boost::vector_property_map<base::Cost> rank(boost::num_vertices(g_));
635
636 try
637 {
638 // Consider using a persistent distance_map if it's slow
639 boost::astar_search(
640 g_, start, [this, goal](Vertex v) { return costHeuristic(v, goal); },
641 boost::predecessor_map(prev)
642 .distance_map(dist)
643 .rank_map(rank)
644 .distance_compare(
645 [this](base::Cost c1, base::Cost c2) { return opt_->isCostBetterThan(c1, c2); })
646 .distance_combine([this](base::Cost c1, base::Cost c2) { return opt_->combineCosts(c1, c2); })
647 .distance_inf(opt_->infiniteCost())
648 .distance_zero(opt_->identityCost())
649 .visitor(AStarGoalVisitor<Vertex>(goal)));
650 }
651 catch (AStarFoundGoal &)
652 {
653 }
654
655 Vertex closeToGoal = start;
656 for (auto vp = vertices(g_); vp.first != vp.second; vp.first++)
657 {
658 // We want to get the distance of each vertex to the goal.
659 // Boost lets us get cost-to-come, cost-to-come+dist-to-goal,
660 // but not just dist-to-goal.
661 ompl::base::Cost dist_to_goal(costHeuristic(*vp.first, goal));
662 if (opt_->isFinite(rank[*vp.first]) && opt_->isCostBetterThan(dist_to_goal, closestVal))
663 {
664 closeToGoal = *vp.first;
665 closestVal = dist_to_goal;
666 approxPathJustStart = false;
667 }
668 }
669 if (closeToGoal != start)
670 {
671 auto p(std::make_shared<PathGeometric>(si_));
672 for (Vertex pos = closeToGoal; prev[pos] != pos; pos = prev[pos])
673 p->append(stateProperty_[pos]);
674 p->append(stateProperty_[start]);
675 p->reverse();
676
677 solution = p;
678 }
679 }
680 }
681 if (approxPathJustStart)
682 {
683 return opt_->infiniteCost();
684 }
685 return closestVal;
686}
687
688ompl::base::PathPtr ompl::geometric::PRM::constructSolution(const Vertex &start, const Vertex &goal)
689{
690 std::lock_guard<std::mutex> _(graphMutex_);
691 boost::vector_property_map<Vertex> prev(boost::num_vertices(g_));
692
693 try
694 {
695 // Consider using a persistent distance_map if it's slow
696 boost::astar_search(
697 g_, start, [this, goal](Vertex v) { return costHeuristic(v, goal); },
698 boost::predecessor_map(prev)
699 .distance_compare([this](base::Cost c1, base::Cost c2) { return opt_->isCostBetterThan(c1, c2); })
700 .distance_combine([this](base::Cost c1, base::Cost c2) { return opt_->combineCosts(c1, c2); })
701 .distance_inf(opt_->infiniteCost())
702 .distance_zero(opt_->identityCost())
703 .visitor(AStarGoalVisitor<Vertex>(goal)));
704 }
705 catch (AStarFoundGoal &)
706 {
707 }
708
709 if (prev[goal] == goal)
710 throw Exception(name_, "Could not find solution path");
711
712 auto p(std::make_shared<PathGeometric>(si_));
713 for (Vertex pos = goal; prev[pos] != pos; pos = prev[pos])
714 p->append(stateProperty_[pos]);
715 p->append(stateProperty_[start]);
716 p->reverse();
717
718 return p;
719}
720
722{
723 Planner::getPlannerData(data);
724
725 // Explicitly add start and goal states:
726 for (unsigned long i : startM_)
727 data.addStartVertex(
728 base::PlannerDataVertex(stateProperty_[i], const_cast<PRM *>(this)->disjointSets_.find_set(i)));
729
730 for (unsigned long i : goalM_)
731 data.addGoalVertex(
732 base::PlannerDataVertex(stateProperty_[i], const_cast<PRM *>(this)->disjointSets_.find_set(i)));
733
734 // Adding edges and all other vertices simultaneously
735 foreach (const Edge e, boost::edges(g_))
736 {
737 const Vertex v1 = boost::source(e, g_);
738 const Vertex v2 = boost::target(e, g_);
740
741 // Add the reverse edge, since we're constructing an undirected roadmap
743
744 // Add tags for the newly added vertices
745 data.tagState(stateProperty_[v1], const_cast<PRM *>(this)->disjointSets_.find_set(v1));
746 data.tagState(stateProperty_[v2], const_cast<PRM *>(this)->disjointSets_.find_set(v2));
747 }
748}
749
751{
752 return opt_->motionCostHeuristic(stateProperty_[u], stateProperty_[v]);
753}
The exception type for ompl.
Definition Exception.h:47
A container that supports probabilistic sampling over weighted data.
Definition PDF.h:49
bool empty() const
Returns whether the PDF contains no data.
Definition PDF.h:263
_T & sample(double r) const
Returns a piece of data from the PDF according to the input sampling value, which must be between 0 a...
Definition PDF.h:132
Element * add(const _T &d, const double w)
Adds a piece of data with a given weight to the PDF. Returns a corresponding Element,...
Definition PDF.h:97
Definition of a cost value. Can represent the cost of a motion or the cost of a state.
Definition Cost.h:48
double value() const
The value of the cost.
Definition Cost.h:56
Abstract definition of a goal region that can be sampled.
Abstract definition of goals.
Definition Goal.h:63
virtual bool isStartGoalPairValid(const State *, const State *) const
Since there can be multiple starting states (and multiple goal states) it is possible certain pairs a...
Definition Goal.h:136
Base class for a vertex in the PlannerData structure. All derived classes must implement the clone an...
Definition PlannerData.h:59
Object containing planner generated vertex and edge data. It is assumed that all vertices are unique,...
bool getEdgeWeight(unsigned int v1, unsigned int v2, Cost *weight) const
Returns the weight of the edge between the given vertex indices. If there exists an edge between v1 a...
bool tagState(const State *st, int tag)
Set the integer tag associated with the given state. If the given state does not exist in a vertex,...
unsigned int addStartVertex(const PlannerDataVertex &v)
Adds the given vertex to the graph data, and marks it as a start vertex. The vertex index is returned...
unsigned int addGoalVertex(const PlannerDataVertex &v)
Adds the given vertex to the graph data, and marks it as a start vertex. The vertex index is returned...
unsigned int numVertices() const
Retrieve the number of vertices in this structure.
unsigned int getEdges(unsigned int v, std::vector< unsigned int > &edgeList) const
Returns a list of the vertex indexes directly connected to vertex with index v (outgoing edges)....
virtual bool addEdge(unsigned int v1, unsigned int v2, const PlannerDataEdge &edge=PlannerDataEdge(), Cost weight=Cost(1.0))
Adds a directed edge between the given vertex indexes. An optional edge structure and weight can be s...
Encapsulate a termination condition for a motion planner. Planners will call operator() to decide whe...
bool isSetup() const
Check if setup() was called for this planner.
Definition Planner.cpp:113
PlannerInputStates pis_
Utility class to extract valid input states.
Definition Planner.h:416
void addPlannerProgressProperty(const std::string &progressPropertyName, const PlannerProgressProperty &prop)
Add a planner progress property called progressPropertyName with a property querying function prop to...
Definition Planner.h:403
PlannerSpecs specs_
The specifications of the planner (its capabilities)
Definition Planner.h:422
ProblemDefinitionPtr pdef_
The user set problem definition.
Definition Planner.h:413
const SpaceInformationPtr & getSpaceInformation() const
Get the space information this planner is using.
Definition Planner.cpp:66
std::string name_
The name of this planner.
Definition Planner.h:419
const std::string & getName() const
Get the name of the planner.
Definition Planner.cpp:56
SpaceInformationPtr si_
The space information for which planning is done.
Definition Planner.h:410
virtual void checkValidity()
Check to see if the planner is in a working state (setup has been called, a goal was set,...
Definition Planner.cpp:106
bool setup_
Flag indicating whether setup() has been called.
Definition Planner.h:433
Definition of an abstract state.
Definition State.h:50
Make the minimal number of connections required to ensure asymptotic optimality.
base::Cost costHeuristic(Vertex u, Vertex v) const
Given two vertices, returns a heuristic on the cost of the path connecting them. This method wraps Op...
Definition PRM.cpp:750
std::mutex graphMutex_
Mutex to guard access to the Graph member (g_)
Definition PRM.h:424
bool starStrategy_
Flag indicating whether the default connection strategy is the Star strategy.
Definition PRM.h:370
ompl::base::Cost constructApproximateSolution(const std::vector< Vertex > &starts, const std::vector< Vertex > &goals, base::PathPtr &solution)
(Assuming that there is always an approximate solution), finds an approximate solution.
Definition PRM.cpp:608
std::vector< Vertex > startM_
Array of start milestones.
Definition PRM.h:385
base::PlannerStatus solve(const base::PlannerTerminationCondition &ptc) override
Function that can solve the motion planning problem. Grows a roadmap using constructRoadmap()....
Definition PRM.cpp:446
bool addedNewSolution() const
Returns the value of the addedNewSolution_ member.
Definition PRM.cpp:441
boost::disjoint_sets< boost::property_map< Graph, boost::vertex_rank_t >::type, boost::property_map< Graph, boost::vertex_predecessor_t >::type > disjointSets_
Data structure that maintains the connected components.
Definition PRM.h:405
void clearQuery() override
Clear the query previously loaded from the ProblemDefinition. Subsequent calls to solve() will reuse ...
Definition PRM.cpp:230
bool sameComponent(Vertex m1, Vertex m2)
Check if two milestones (m1 and m2) are part of the same connected component. This is not a const fun...
Definition PRM.cpp:603
base::OptimizationObjectivePtr opt_
Objective cost function for PRM graph edges.
Definition PRM.h:427
bool addedNewSolution_
A flag indicating that a solution has been added during solve()
Definition PRM.h:421
base::Cost bestCost_
Best cost found so far by algorithm.
Definition PRM.h:434
void freeMemory()
Free all the memory allocated by the planner.
Definition PRM.cpp:251
base::StateSamplerPtr simpleSampler_
Sampler user for generating random in the state space.
Definition PRM.h:376
RoadmapNeighbors nn_
Nearest neighbors data structure.
Definition PRM.h:379
boost::graph_traits< Graph >::edge_descriptor Edge
The type for an edge in the roadmap.
Definition PRM.h:127
boost::graph_traits< Graph >::vertex_descriptor Vertex
The type for a vertex in the roadmap.
Definition PRM.h:125
PRM(const base::SpaceInformationPtr &si, bool starStrategy=false)
Constructor.
Definition PRM.cpp:72
void checkForSolution(const base::PlannerTerminationCondition &ptc, base::PathPtr &solution)
Definition PRM.cpp:379
void expandRoadmap(double expandTime)
Attempt to connect disjoint components in the roadmap using random bouncing motions (the PRM expansio...
Definition PRM.cpp:258
double distanceFunction(const Vertex a, const Vertex b) const
Compute distance between two milestones (this is simply distance between the states of the milestones...
Definition PRM.h:345
bool maybeConstructSolution(const std::vector< Vertex > &starts, const std::vector< Vertex > &goals, base::PathPtr &solution)
Check if there exists a solution, i.e., there exists a pair of milestones such that the first is in s...
Definition PRM.cpp:400
ConnectionStrategy connectionStrategy_
Function that returns the milestones to attempt connections with.
Definition PRM.h:408
base::PathPtr constructSolution(const Vertex &start, const Vertex &goal)
Given two milestones from the same connected component, construct a path connecting them and set it a...
Definition PRM.cpp:688
void uniteComponents(Vertex m1, Vertex m2)
Make two milestones (m1 and m2) be part of the same connected component. The component with fewer ele...
Definition PRM.cpp:598
void constructRoadmap(const base::PlannerTerminationCondition &ptc)
While the termination condition allows, this function will construct the roadmap (using growRoadmap()...
Definition PRM.cpp:530
boost::property_map< Graph, boost::edge_weight_t >::type weightProperty_
Access to the weights of each Edge.
Definition PRM.h:401
boost::property_map< Graph, vertex_successful_connection_attempts_t >::type successfulConnectionAttemptsProperty_
Access to the number of successful connection attempts for a vertex.
Definition PRM.h:398
unsigned long int iterations_
Number of iterations the algorithm performed.
Definition PRM.h:432
void growRoadmap(double growTime)
If the user desires, the roadmap can be improved for the given time (seconds). The solve() method wil...
Definition PRM.cpp:339
Graph g_
Connectivity graph.
Definition PRM.h:382
Vertex addMilestone(base::State *state)
Construct a milestone for a given state (state), store it in the nearest neighbors data structure and...
Definition PRM.cpp:562
unsigned long int milestoneCount() const
Return the number of milestones currently in the graph.
Definition PRM.h:276
void setup() override
Perform extra configuration steps, if needed. This call will also issue a call to ompl::base::SpaceIn...
Definition PRM.cpp:155
bool userSetConnectionStrategy_
Flag indicating whether the employed connection strategy was set by the user (or defaults are assumed...
Definition PRM.h:415
void clear() override
Clear all internal datastructures. Planner settings are not affected. Subsequent calls to solve() wil...
Definition PRM.cpp:237
void setMaxNearestNeighbors(unsigned int k)
Convenience function that sets the connection strategy to the default one with k nearest neighbors.
Definition PRM.cpp:193
void getPlannerData(base::PlannerData &data) const override
Get information about the current run of the motion planner. Repeated calls to this function will upd...
Definition PRM.cpp:721
std::vector< Vertex > goalM_
Array of goal milestones.
Definition PRM.h:388
unsigned int getMaxNearestNeighbors() const
return the maximum number of nearest neighbors to connect a sample to
Definition PRM.cpp:210
base::ValidStateSamplerPtr sampler_
Sampler user for generating valid samples in the state space.
Definition PRM.h:373
void setDefaultConnectionStrategy()
Definition PRM.cpp:216
RNG rng_
Random number generator.
Definition PRM.h:418
boost::property_map< Graph, vertex_state_t >::type stateProperty_
Access to the internal base::state at each Vertex.
Definition PRM.h:391
boost::property_map< Graph, vertex_total_connection_attempts_t >::type totalConnectionAttemptsProperty_
Access to the number of total connection attempts for a vertex.
Definition PRM.h:394
ConnectionFilter connectionFilter_
Function that can reject a milestone connection.
Definition PRM.h:411
static NearestNeighbors< _T > * getDefaultNearestNeighbors(const base::Planner *planner)
Select a default nearest neighbor datastructure for the given space.
Definition SelfConfig.h:106
#define OMPL_INFORM(fmt,...)
Log a formatted information string.
Definition Console.h:68
#define OMPL_ERROR(fmt,...)
Log a formatted error string.
Definition Console.h:64
This namespace contains sampling based planning routines shared by both planning under geometric cons...
@ GOAL_SAMPLEABLE_REGION
This bit is set if casting to sampleable goal regions (ompl::base::GoalSampleableRegion) is possible.
Definition GoalTypes.h:56
PlannerTerminationCondition plannerOrTerminationCondition(const PlannerTerminationCondition &c1, const PlannerTerminationCondition &c2)
Combine two termination conditions into one. If either termination condition returns true,...
PlannerTerminationCondition timedPlannerTerminationCondition(double duration)
Return a termination condition that will become true duration seconds in the future (wall-time)
This namespace includes magic constants used in various places in OMPL.
Definition Constraint.h:52
static const unsigned int FIND_VALID_STATE_ATTEMPTS_WITHOUT_TERMINATION_CHECK
Maximum number of sampling attempts to find a valid state, without checking whether the allowed time ...
static const double ROADMAP_BUILD_TIME
The time in seconds for a single roadmap building operation (dt)
Definition PRM.cpp:64
static const unsigned int DEFAULT_NEAREST_NEIGHBORS
The number of nearest neighbors to consider by default in the construction of the PRM roadmap.
Definition PRM.cpp:68
static const unsigned int MAX_RANDOM_BOUNCE_STEPS
The number of steps to take for a random bounce motion generated as part of the expansion step of PRM...
Definition PRM.cpp:61
Main namespace. Contains everything in this library.
Representation of a solution to a planning problem.
void setPlannerName(const std::string &name)
Set the name of the planner used to compute this solution.
void setOptimized(const OptimizationObjectivePtr &opt, Cost cost, bool meetsObjective)
Set the optimization objective used to optimize this solution, the cost of the solution and whether i...
A class to store the exit status of Planner::solve()
@ INVALID_START
Invalid start state or no start state specified.
@ EXACT_SOLUTION
The planner found an exact solution.
@ INVALID_GOAL
Invalid goal state.
@ UNRECOGNIZED_GOAL_TYPE
The goal is of a type that a planner does not recognize.
@ APPROXIMATE_SOLUTION
The planner found an approximate solution.
@ TIMEOUT
The planner failed to find a solution.