## The Ariadne's clew algorithm (1996)

### Cached

### Download Links

- [ftp.inrialpes.fr]
- [www-leibniz.imag.fr]
- [ftp.imag.fr]
- DBLP

### Other Repositories/Bibliography

Citations: | 73 - 3 self |

### BibTeX

@MISC{Mazer96theariadne's,

author = {Emmanuel Mazer and Juan Manuel Ahuactzin and El-Ghazali Talbi and Pierre Bessiere},

title = {The Ariadne's clew algorithm},

year = {1996}

}

### Years of Citing Articles

### OpenURL

### Abstract

We present a general planning strategy to plan the motions of an agent having to explore a continuous state space in order to reach one or several goals. We propose a practical method to implement this technique based on a genetic algorithm and we illustrate the approach on the problem of controlling a mobile robot moving in a maze and looking for several items. Finally, we show that this planning strategy may serve as a possible control structure for an autonomous system. Problem Solving and planning 1 Introduction This study was motivated by our previous work on robot motion planning using a parallel genetic algorithm [8]. The planner we have design and implemented on a parallel machine is capable of planning collision free paths for a mobile robot placed among obstacles. The main advantage of this planner is its speed, it can plan complex paths such as the two paths represented in figure 1 in less than 0.5 second on a parallel machine with 64 Transputers. As a consequence it can b...

### Citations

2818 | Adaptation in Natural and Artificial Systems - Holland - 1975 |

888 | Probabilistic roadmaps for path planning in high-dimensional configuration space
- Svestka, Overmars
- 1996
(Show Context)
Citation Context ...s capable of such performance. To the best of our knowledge, currently implemented planners would take a number of seconds (ten) to place a set of landmarks on a 2d example for a robot with five dof (=-=Kavraki et al., 1996-=-). Despite the fact that finding a general purpose planning technique for real industrial application is a very difficult problem, we believe that the Ariadne's clew algorithm provides an effective ap... |

479 |
The complexity of robot motion planning
- Canny
- 1988
(Show Context)
Citation Context ...sions with any obstacles in the environment. A simple version of the problem, that of planning the motion of a point robot among 3-dimensional polyhedral obstacles, has been proved to be np-complete (=-=Canny, 1988-=-). Generally speaking, the complexity of the problem is exponential in the number of degrees of freedom (dof) of the robot, and polynomial in the number of obstacles in the environment. Consequently, ... |

335 |
Robot Motion Planning: A Distributed Representation Approach
- Barraquand, Latombe
- 1991
(Show Context)
Citation Context ...of faces which are far out of reach of global methods. 2.3 Path planning with randomized techniques and landmarks The stochastic or random approach was first introduced by Barraquand and Latombe [15] =-=[14]-=-, and used later by Overmars [17][16] and more recently by Kavraki[3]. The idea of this type of algorithm is to build a graph in the configuration space. The graph is obtained incrementally as follow ... |

155 | Gross Motion Planning - A Survey - HWANG, AHUJA - 1992 |

118 | Artificial life and real robots - Brooks - 1992 |

94 |
Solving the Find-Path Problem by Good Representation of Space
- Brooks
- 1982
(Show Context)
Citation Context ...t arm, the problem is to find a set of motions that will lead the robot to move between the two positions without colliding with the obstacles. To drive the robot amidst the obstacles, early methods (=-=Brooks, 1983) directly-=- used the 3d cad models of the robot and of the obstacles to find a solution, i.e., they considered the "operational 3d space". In this space, the path planning problem consists of finding t... |

91 |
Parallel genetic algorithms for a hypercube
- Tanese
- 1987
(Show Context)
Citation Context ... operation. 4.1.2 Principle of the Parallel Genetic Algorithm (pga) There are many parallel versions of genetic algorithms: the standard parallel version (Robertson, 1987), the decomposition version (=-=Tanese, 1987-=-) and the massively parallel version (Talbi, 1993). We chose this last method. The main idea is to allocate one element of the population for each processor so that steps 1, 3, and 4 can be executed i... |

87 | A Bootom-Up Mechanism for Behavior Selection in an Artificial Creature," in From Animals to Animats - Maes - 1991 |

76 | A Simple Motion-Planning Algorithm for General Robot Manipulators
- LOZANO-PEREZ
- 1987
(Show Context)
Citation Context ...sts of finding the movements of a complex 3d structure (the robot) in a cluttered 3d space. A major advance was to express the problem in another space known as the configuration space, denoted by C (=-=Lozano-P'erez, 1987-=-). In this space, the position (or configuration) of a robot is completely determined by a single point having n independent parameters as coordinates. The positions that are not physically legal (bec... |

69 |
A parallel genetic algorithm
- Pettey, Leuze, et al.
- 1987
(Show Context)
Citation Context ... newly produced element. 5.1.2 Principle of the parallel genetic algorithm (PGA) There are many parallel versions of genetic algorithms : the parallel standard version [31], the decomposition version =-=[32]-=-[33] and the massively parallel version [29]. We chose this last method. The principle is to place one element of the population per processor so the steps 1, 3 and 4 can be executed in parallel. Furt... |

59 | Latombe, “Landmark-based robot navigation - Lazanas, C - 1992 |

57 | Challenges for Complete Creature Architectures. In From Animals to Animats - Brooks - 1990 |

53 | A random approach to motion planning - Overmars - 1992 |

51 | Evolution and co-evolution of computer programs to control independently acting agents - Koza - 1991 |

45 | A local based approach for path planning of manipulators with a high number of degrees of freedom - Faverjon, Tournassoud - 1987 |

44 | Evolution of subsumption using genetic programming - Koza - 1992 |

43 | A genetic algorithm for bin packing and line balancing - Falkenauer, Delchambre - 1992 |

39 | Learning behavior networks from experience. In Toward a Practice of Autonomous Systems - Maes - 1992 |

37 |
Parallel implementation of genetic algorithms on a classifier system
- Robertson
- 1987
(Show Context)
Citation Context ...ss-over NEW ELEMENTS Figure 9: The cross-over operation. 4.1.2 Principle of the Parallel Genetic Algorithm (pga) There are many parallel versions of genetic algorithms: the standard parallel version (=-=Robertson, 1987-=-), the decomposition version (Tanese, 1987) and the massively parallel version (Talbi, 1993). We chose this last method. The main idea is to allocate one element of the population for each processor s... |

37 | Robot motion planning with uncertainty in control and sensing - Latombe, Lazanas, et al. - 1991 |

37 | A geometric approach to error detection and recovery for robot motion planning with uncertainty. Artificial Intell - Donald - 1988 |

31 |
A parallel genetic algorithm for the graph partitioning problem
- Talbi, Bessière
- 1991
(Show Context)
Citation Context ...mplement on parallel architecture unlike number of the other optimization technics [4]. We previously developped a parallel genetic algorithm (PGA) and we already had an important experience using it =-=[7, 8, 6]-=-. 3. PGA, unlike most parallel programs, shows linear speed-up (when you double the number of processors you divide by 2 the needded amount of time) and even super linear speed-up under certain circun... |

31 |
SANDROS: A motion planner with performance proportional to task di - culty
- Chen, Hwang
- 1992
(Show Context)
Citation Context ...ots with a high number of degrees of freedom and they have been shown to work fast in relatively complex and natural environment. Other methods using landmarks have been devised. For example, SANDROS =-=[13] introduce-=-d by Chen and Hwang makes use of landmarks to approximate the free space. This approach is simliar to the "hierarchical planning" approach used in AI : should the method fail to reach a goal... |

28 | Using genetic algorithms for robot motion planning - Ahuactzin, Mazer, et al. - 1992 |

24 | Maniopulation planning for redundant robots: A practical approach - Ahuactzin, Mazer - 1998 |

19 | Genetic algorithms and machine learning - Goldberg, Holland - 1988 |

17 |
Analogous crossover
- Davidor
- 1989
(Show Context)
Citation Context ...ion: Produce a new element from each pair using "genetic" operators. 4. Replacement: Replace the elements of the starting population by better new elements produced in step 3. Many genetic o=-=perators (Davidor, 1989-=-) are available. However, the more commonly used are the mutation and the cross-over operators. The mutation operator consists of randomly flipping some bits of an element of the population. The cross... |

17 | A Genetic Algorithm for Job Shop - Falkenauer, Bouffouix - 1991 |

16 | Continuous-curvature path planning for car-like vehicles - Scheuer, Fraichard - 1997 |

15 |
Hwang and Narendra Ahuja. Gross motion planning – A survey
- Yong
- 1992
(Show Context)
Citation Context ...lanning problem Many versions of the path planning problem exist. An exhaustive classification of these problems and of the methods developed to solve them can be found in a survey by Hwang and Ahuja =-=[10]-=-. We chose to illustrate our algorithm by considering a particular case. A robot arm is placed among a set of obstacles. Given an initial and a final position of the robot arm find a set of commands w... |

15 | Atlanta). Automatic planning of fine motions: correctness and completeness - Mason - 1984 |

13 |
Handbook of genetic algorithms
- Lawrence
- 1991
(Show Context)
Citation Context ...orithm Genetic algorithms are stochastic optimization techniques introduced by Holland (1975) twenty years ago. They are used in a large variety of domains including robotics (Ahuactzin et al., 1992; =-=Lawrence, 1991-=-; Falkenauer & Bouffouix, 1991; Falkenauer & Delchambre, 1992; Meygret & Levine, 1992) because they are easy to implement and do not require algebraic expression for the function to be optimized. 4.1.... |

13 | Une Théorie Asymptotique des Algorithmes Génétiques - Cerf - 1994 |

11 | Incremental roadmaps and global path planning in evolving industrial environments - McLean, Mazon - 1996 |

11 |
Planning continuous-curvature paths for car-like robots
- Fraichard
- 1996
(Show Context)
Citation Context ... the evaluation functions of the 2 optimization problems. Several planners have been implemented this way: a fine motion planner [37], two motion planner for holonomic and non holonomic mobile robots =-=[38]-=-, a reorientation planner for an articulated handxxx, a planner for grasping and regrasping xxx and a planner for a robotic arm placed in the steam generator of a nuclear plant [39]. Adapting the algo... |

11 | Goldberg - Genetic Algorithms in Search, optimization and Machine Learning - David |

10 | An Overview of Kali: a System to Program and Control Cooperative Manipulators - Hayward, Daneshmend, et al. - 1989 |

8 |
Motion planning for re-orientation using finger tracking
- Gupta
- 1995
(Show Context)
Citation Context ...ne motion planner (De la Rosa, Laugier, & Najera, 1996), two motion planners for holonomic and non-holonomic mobile robots (Scheuer & Fraichard, 1997), a reorientation planner for an articulated hand(=-=Gupta, 1995-=-), a planner for grasping and regrasping (Ahuactzin, Gupta, & Mazer, 1998), and a planner for a robotic arm placed in the steam generator of a nuclear plant (McLean & Mazon, 1996). Adapting the algori... |

7 | Robust path planning in the plane - Rosa, Laugier, et al. - 1996 |

7 | Evolutionary navigator for a mobile robot - Lin, Xiao, et al. - 1994 |

5 |
Allocation de processus sur les architectures parall`eles `a m'emoire distribu'ee
- Talbi
- 1993
(Show Context)
Citation Context ...uactzin, & Mazer, 1996), are very easy to implement on parallel architectures. We have previously developed a parallel genetic algorithm (pga) and we have already had significant experience using it (=-=Talbi, 1993-=-). 3. pga, unlike most parallel programs, shows linear speed-up (when you double the number of processors you reduce the computation time by half) and even super-linear speed-up under certain circumst... |

4 |
Latombe, “A monte-carlo algorithm for path planning with many degrees of freedom
- Barraquand, C
- 1990
(Show Context)
Citation Context ...ands of faces which are far out of reach of global methods. 2.3 Path planning with randomized techniques and landmarks The stochastic or random approach was first introduced by Barraquand and Latombe =-=[15]-=- [14], and used later by Overmars [17][16] and more recently by Kavraki[3]. The idea of this type of algorithm is to build a graph in the configuration space. The graph is obtained incrementally as fo... |

3 | A parallel genetic algorithm applied to the mapping problem - Talbi, Bessi`ere - 1991 |

3 | Zhenping Guo: Motion Planning for Many Degrees of Freedom Sequential Search With Backtracking - Gupta - 1992 |

3 | Latombe: Robot Motion Planning, Ed - Jean-Claude - 1991 |

2 |
Overmars: A random approach to motion planning
- Mark
- 1992
(Show Context)
Citation Context ...f global methods. 2.3 Path planning with randomized techniques and landmarks The stochastic or random approach was first introduced by Barraquand and Latombe [15] [14], and used later by Overmars [17]=-=[16]-=- and more recently by Kavraki[3]. The idea of this type of algorithm is to build a graph in the configuration space. The graph is obtained incrementally as follow : a local planner is used to try to r... |

1 | The Ariadne's clew algorithm - Barraquand - 1990 |

1 | Un Algorithme G'en'etique Parall`elle pour l'Optimisation. Technique et - Bessi`ere, Talbi, et al. - 1996 |

1 |
Contribution `a la Planification de Trajectoires. Rapport de Recherche EDF-DER HP-28/96/026, Direction des ' Etudes et Recherches EDF
- Ferbach
- 1996
(Show Context)
Citation Context ...nding positions P (q a ) and P (q f ) in the operational space. This curve is one solution to this particular path planning problem. A recent trend in the field is to consider the "trajectory spa=-=ce" (Ferbach, 1996-=-) where a whole path is represented by a single point. The coordinates of this point are the values of the parameters defining the successive movements of the robot. For instance, the list of successi... |