Browse ORBi by ORBi project

- Background
- Content
- Benefits and challenges
- Legal aspects
- Functions and services
- Team
- Help and tutorials

Effective segmentation of green vegetation for resource-constrained real-time applications Krishna Moorthy Parvathi, Sruthi Moorthy ; Boigelot, Bernard ; Mercatoris, Benoît Scientific conference (2015, July 15) This paper describes an improved algorithm for segmentation of green vegetation under uncontrolled illumination conditions and also suitable for resource-constrained real-time applications. The proposed ... [more ▼] This paper describes an improved algorithm for segmentation of green vegetation under uncontrolled illumination conditions and also suitable for resource-constrained real-time applications. The proposed algorithm uses a naïve Bayesian model to effectively combine various manually extracted features from two different color spaces namely RGB and HSV. The evaluation of 100 images indicated the better performance of the proposed algorithm than the vegetation index-based methods with comparable execution time. Moreover, the proposed algorithm performed better than the state-of-the-art EASA-based algorithms in terms of processing time and memory usage. [less ▲] Detailed reference viewed: 83 (11 ULg)Efficient Path Interpolation and Speed Profile Computation for Nonholonomic Mobile Robots Lens, Stéphane ; Boigelot, Bernard E-print/Working paper (2015) This paper studies path synthesis for nonholonomic mobile robots moving in two-dimensional space. We first address the problem of interpolating paths expressed as sequences of straight line segments, such ... [more ▼] This paper studies path synthesis for nonholonomic mobile robots moving in two-dimensional space. We first address the problem of interpolating paths expressed as sequences of straight line segments, such as those produced by some planning algorithms, into smooth curves that can be followed without stopping. Our solution has the advantage of being simpler than other existing approaches, and has a low computational cost that allows a real-time implementation. It produces discretized paths on which curvature and variation of curvature are bounded at all points, and preserves obstacle clearance. Then, we consider the problem of computing a time-optimal speed profile for such paths. We introduce an algorithm that solves this problem in linear time, and that is able to take into account a broader class of physical constraints than other solutions. Our contributions have been implemented and evaluated in the framework of the Eurobot contest. [less ▲] Detailed reference viewed: 37 (11 ULg)Abstract: A36.00003 : Interaction between electrically charged droplets in microgravity Brandenbourger, Martin ; Caps, Hervé ; et al Conference (2015) Detailed reference viewed: 18 (5 ULg)Efficient Path Planning for Nonholonomic Mobile Robots Lens, Stéphane ; Boigelot, Bernard Report (2014) This work addresses path planning for nonholonomic robots moving in two-dimensional space. The problem consists in computing a sequence of line segments that leads from the current configuration of the ... [more ▼] This work addresses path planning for nonholonomic robots moving in two-dimensional space. The problem consists in computing a sequence of line segments that leads from the current configuration of the robot to a target location, while avoiding a given set of obstacles. We describe a planning algorithm that has the advantage of being very efficient, requiring less of one millisecond of CPU time for the case studies that we have considered, and produces short paths. Our method relies on a search in a Voronoi graph that characterizes the possible ways of moving around obstacles, followed by a string-pulling procedure aimed at improving the resulting path. [less ▲] Detailed reference viewed: 45 (9 ULg)A vision-based autonomous inter-row weeder Krishna Moorthy Parvathi, Sruthi Moorthy ; Detry, Renaud ; Boigelot, Bernard et al Conference (2014, March 05) Autonomous robotic weed destruction plays a significant role in crop production as it automates one of the few unmechanized and drudging tasks of agriculture i.e. manual weed destruction. Robotic ... [more ▼] Autonomous robotic weed destruction plays a significant role in crop production as it automates one of the few unmechanized and drudging tasks of agriculture i.e. manual weed destruction. Robotic technology also contributes to long-term sustainability with both economic and environmental benefits, by minimising the current dependency on chemicals. The aim of this study is to design a small low-cost versatile robot allowing the destruction of weeds that lie between the crop rows by navigating in the field autonomously and using a minimum of a priori information of the field. For the robot to navigate autonomously, necessary and sufficient information can be supplied by a machine vision system. One important issue with the application of machine vision is to develop a system that recognises the crop rows accurately and robustly which is tolerant to problems such as crops at varying growth stages, poor illumination conditions, missing crops, high weed pressure, etc. Aiming at accurate and robust real-time guidance of autonomous robot through the field, the plethora of image processing algorithms like Ostu’s threshold method and hough transform will be explored for two main processes namely the image segmentation and crop row detection respectively. In order to overcome the issue of large variabilities encountered in agriculture such as varying weather conditions, intelligent stochastic data fusion and machine learning algorithms will be used to combine data from heterogeneous sensors. Besides crop row detection, other major challenges foreseen are: mapping the unknown geometry of the field, high-level planning of efficient and complete coverage of the field, controlling the low-level op- erations of the robot, and ensuring security. Specialised sensors such as GPS will be considered to generate the map of the field enabling Simultaneous Localisation And Mapping (SLAM) in real time on a mobile platform. The generated map will be exploited along with the sensorial in- formation from crop row detection to efficiently plan and execute the guidance of the robot au- tonomously in the field, thereby enabling weed elimination. [less ▲] Detailed reference viewed: 203 (17 ULg)Robot weed killers - no pain more gain Krishna Moorthy Parvathi, Sruthi Moorthy ; Mercatoris, Benoît ; Boigelot, Bernard Poster (2014, February 07) Weed destruction plays a significant role in crop production, and its automation has both economic and environmental benefits by minimizing the usage of chemicals in the fields. Our aim is to design a ... [more ▼] Weed destruction plays a significant role in crop production, and its automation has both economic and environmental benefits by minimizing the usage of chemicals in the fields. Our aim is to design a small low-cost versatile robot allowing the destruction of weeds that lie between the crop rows by navigating in the field autonomously. Major challenges foreseen are: mapping the unknown geometry of the field, high-level planning of efficient and complete coverage of the field, and controlling the low-level operations of the robot. Traditionally, sensors like odometer have been used for localisation of robots but without much success in real-world scenarios. Specialized sensors like cameras will therefore be investigated and the plethora of image recognition algorithms will be explored and fine-tuned to enable Simultaneous Localisation And Mapping (SLAM) even on resource constrained robotic platforms. Vision-based localisation is not always viable because of the varying weather conditions of the environment and to overcome that, intelligent stochastic data fusion and machine learning algorithms will be utilized to combine data from heterogenous sensor. The image sensors for localisation will be re-used to differentiate crop rows from the weeds, which are cut when they grow. Finally, logics and reinforcement learning techniques will be explored, to exploit the generated map of the field and other sensorial information, to efficiently plan and execute weed elimination. [less ▲] Detailed reference viewed: 90 (8 ULg)Acceleration of Affine Hybrid Transformations Boigelot, Bernard ; ; Mainz, Isabelle in Lecture Notes in Computer Science (2014), 8837 This work addresses the computation of the set of reachable configurations of linear hybrid automata. The approach relies on symbolic state-space exploration, using acceleration in order to speed up the ... [more ▼] This work addresses the computation of the set of reachable configurations of linear hybrid automata. The approach relies on symbolic state-space exploration, using acceleration in order to speed up the computation and to make it terminate for a broad class of systems. Our contribution is an original method for accelerating the control cycles of linear hybrid automata, i.e., to compute their unbounded repeated effect. The idea consists in analyzing the data transformations that label these cycles, by reasoning about the geometrical features of the corresponding system of linear constraints. This approach is complete over Multiple Counters Systems (MCS), and is able to accelerate hybrid transformations that are out of scope of existing techniques. [less ▲] Detailed reference viewed: 55 (15 ULg)Lipschitz robust control from off-policy trajectories Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al in Proceedings of the 53rd IEEE Conference on Decision and Control (IEEE CDC 2014) (2014) We study the minmax optimization problem introduced in [Fonteneau et al. (2011), ``Towards min max reinforcement learning'', Springer CCIS, vol. 129, pp. 61-77] for computing control policies for batch ... [more ▼] We study the minmax optimization problem introduced in [Fonteneau et al. (2011), ``Towards min max reinforcement learning'', Springer CCIS, vol. 129, pp. 61-77] for computing control policies for batch mode reinforcement learning in a deterministic setting with fixed, finite optimization horizon. First, we state that the $\min$ part of this problem is NP-hard. We then provide two relaxation schemes. The first relaxation scheme works by dropping some constraints in order to obtain a problem that is solvable in polynomial time. The second relaxation scheme, based on a Lagrangian relaxation where all constraints are dualized, can also be solved in polynomial time. We theoretically show that both relaxation schemes provide better results than those given in [Fonteneau et al. (2011)] [less ▲] Detailed reference viewed: 57 (10 ULg)Généralisation Min Max pour l'Apprentissage par Renforcement Batch et Déterministe : Relaxations pour le Cas Général T Etapes Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al in 8èmes Journées Francophones de Planification, Décision et Apprentissage pour la conduite de systèmes (JFPDA'13) (2013) Cet article aborde le problème de généralisation minmax dans le cadre de l'apprentissage par renforcement batch et déterministe. Le problème a été originellement introduit par [Fonteneau, 2011], et il a ... [more ▼] Cet article aborde le problème de généralisation minmax dans le cadre de l'apprentissage par renforcement batch et déterministe. Le problème a été originellement introduit par [Fonteneau, 2011], et il a déjà été montré qu'il est NP-dur. Deux schémas de relaxation pour le cas deux étapes ont été présentés aux JFPDA'12, et ce papier présente une généralisation de ces schémas au cas T étapes. Le premier schéma fonctionne en éliminant des contraintes afin d'obtenir un problème soluble en temps polynomial. Le deuxième schéma est une relaxation lagrangienne conduisant également à un problème soluble en temps polynomial. On montre théoriquement que ces deux schémas permettent d'obtenir de meilleurs résultats que ceux proposés par [Fonteneau, 2011]. [less ▲] Detailed reference viewed: 45 (7 ULg)Min max generalization for deterministic batch mode reinforcement learning: relaxation schemes Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al in SIAM Journal on Control & Optimization (2013), 51(5), 33553385 We study the min max optimization problem introduced in Fonteneau et al. [Towards min max reinforcement learning, ICAART 2010, Springer, Heidelberg, 2011, pp. 61–77] for computing policies for batch mode ... [more ▼] We study the min max optimization problem introduced in Fonteneau et al. [Towards min max reinforcement learning, ICAART 2010, Springer, Heidelberg, 2011, pp. 61–77] for computing policies for batch mode reinforcement learning in a deterministic setting with ﬁxed, ﬁnite time horizon. First, we show that the min part of this problem is NP-hard. We then provide two relaxation schemes. The ﬁrst relaxation scheme works by dropping some constraints in order to obtain a problem that is solvable in polynomial time. The second relaxation scheme, based on a Lagrangian relaxation where all constraints are dualized, can also be solved in polynomial time. We also theoretically prove and empirically illustrate that both relaxation schemes provide better results than those given in [Fonteneau et al., 2011, as cited above]. [less ▲] Detailed reference viewed: 54 (15 ULg)Généralisation min max pour l'apprentissage par renforcement batch et déterministe : schémas de relaxation Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al in Septièmes Journées Francophones de Planification, Décision et Apprentissage pour la conduite de systèmes (JFPDA 2012) (2012, May) On s’intéresse au problème de généralisation min max dans le cadre de l’apprentissage par renforcement batch et déterministe. Le problème a été originellement introduit par Fonteneau et al. (2011). Dans ... [more ▼] On s’intéresse au problème de généralisation min max dans le cadre de l’apprentissage par renforcement batch et déterministe. Le problème a été originellement introduit par Fonteneau et al. (2011). Dans un premier temps, on montre que le problème est NP-dur. Dans le cas où l’horizon d’optimisation vaut 2, on développe deux schémas de relaxation. Le premier schéma fonctionne en éliminant des contraintes de telle sorte qu’on obtienne un problème soluble en temps polynomial. Le deuxième schéma est une relaxation Lagrangienne conduisant à un problème conique-quadratique. On montre théoriquement et empiriquement que ces deux schémas permettent d’obtenir de meilleurs résultats que ceux proposés par Fonteneau et al. (2011). [less ▲] Detailed reference viewed: 77 (10 ULg)Automata-Based Symbolic Representations of Polyhedra Boigelot, Bernard ; Brusten, Julien ; Degbomont, Jean-François in Lecture Notes in Computer Science (2012, March), 7183 Detailed reference viewed: 67 (23 ULg)Min max generalization for two-stage deterministic batch mode reinforcement learning: relaxation schemes Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al Report (2012) Detailed reference viewed: 37 (4 ULg)Domain-specific regular acceleration Boigelot, Bernard in International Journal on Software Tools for Technology Transfer (2012), 14(2), 193-206 The regular model-checking approach is a set of techniques aimed at exploring symbolically infinite state spaces. These techniques proceed by representing sets of configurations of the system under ... [more ▼] The regular model-checking approach is a set of techniques aimed at exploring symbolically infinite state spaces. These techniques proceed by representing sets of configurations of the system under analysis by regular languages, and the transition relation between these configurations by a transformation over such languages. The set of reachable configurations can then be computed by repeatedly applying the transition relation, starting from a representation of the initial set of configurations, until a fixed point is reached. In order for this computation to terminate, it is generally needed to introduce so-called acceleration operators, the purpose of which is to explore in one computation step infinitely many paths in the transition graph of the system. A simple form of acceleration operator is one that is associated to a cycle in the transition graph, computing the set of states that can be obtained by following this cycle arbitrarily many times. The computation of acceleration operators is strongly dependent on the type of the data values that are manipulated by the system, and on the symbolic representation chosen for handling sets of such values. In this survey, we describe acceleration operators suited for the regular state-space exploration of systems relying on FIFO communication channels, as well as those based on integer and real variables. [less ▲] Detailed reference viewed: 29 (8 ULg)Relaxation schemes for min max generalization in deterministic batch mode reinforcement learning Fonteneau, Raphaël ; Ernst, Damien ; Boigelot, Bernard et al in 4th International NIPS Workshop on Optimization for Machine Learning (OPT 2011) (2011, December) We study the min max optimization problem introduced in [Fonteneau, 2011] for computing policies for batch mode reinforcement learning in a deterministic setting. This problem is NP-hard. We focus on the ... [more ▼] We study the min max optimization problem introduced in [Fonteneau, 2011] for computing policies for batch mode reinforcement learning in a deterministic setting. This problem is NP-hard. We focus on the two-stage case for which we provide two relaxation schemes. The first relaxation scheme works by dropping some constraints in order to obtain a problem that is solvable in polynomial time. The second relaxation scheme, based on a Lagrangian relaxation where all constraints are dualized, leads to a conic quadratic programming problem. Both relaxation schemes are shown to provide better results than those given in [Fonteneau, 2011]. [less ▲] Detailed reference viewed: 148 (10 ULg)On the Sets of Real Numbers Recognized by Finite Automata in Multiple Bases Boigelot, Bernard ; Brusten, Julien ; in Logical Methods in Computer Science (2010), 6(1), 1-17 This article studies the expressive power of finite automata recognizing sets of real numbers encoded in positional notation. We consider Muller automata as well as the restricted class of weak ... [more ▼] This article studies the expressive power of finite automata recognizing sets of real numbers encoded in positional notation. We consider Muller automata as well as the restricted class of weak deterministic automata, used as symbolic set representations in actual applications. In previous work, it has been established that the sets of numbers that are recognizable by weak deterministic automata in two bases that do not share the same set of prime factors are exactly those that are definable in the first order additive theory of real and integer numbers. This result extends Cobham's theorem, which characterizes the sets of integer numbers that are recognizable by finite automata in multiple bases. In this article, we first generalize this result to multiplicatively independent bases, which brings it closer to the original statement of Cobham's theorem. Then, we study the sets of reals recognizable by Muller automata in two bases. We show with a counterexample that, in this setting, Cobham's theorem does not generalize to multiplicatively independent bases. Finally, we prove that the sets of reals that are recognizable by Muller automata in two bases that do not share the same set of prime factors are exactly those definable in the first order additive theory of real and integer numbers. These sets are thus also recognizable by weak deterministic automata. This result leads to a precise characterization of the sets of real numbers that are recognizable in multiple bases, and provides a theoretical justification to the use of weak automata as symbolic representations of sets. [less ▲] Detailed reference viewed: 60 (19 ULg)Implicit Real Vector Automata Boigelot, Bernard ; Brusten, Julien ; Degbomont, Jean-François in Electronic Proceedings in Theoretical Computer Science [=EPTCS] (2010), 39 This paper addresses the symbolic representation of non-convex real polyhedra, i.e., sets of real vectors satisfying arbitrary Boolean combinations of linear constraints. We develop an original data ... [more ▼] This paper addresses the symbolic representation of non-convex real polyhedra, i.e., sets of real vectors satisfying arbitrary Boolean combinations of linear constraints. We develop an original data structure for representing such sets, based on an implicit and concise encoding of a known structure, the Real Vector Automaton. The resulting formalism provides a canonical representation of polyhedra, is closed under Boolean operators, and admits an efficient decision procedure for testing the membership of a vector. [less ▲] Detailed reference viewed: 103 (41 ULg)A Generalization of Semenov's Theorem to Automata over Real Numbers Boigelot, Bernard ; Brusten, Julien ; in Lecture Notes in Artificial Intelligence (2009), 5663 Detailed reference viewed: 82 (37 ULg)Partial Projection of Sets Represented by Finite Automata, with Application to State-Space Visualization Boigelot, Bernard ; Degbomont, Jean-François in Lecture Notes in Computer Science (2009), 5457 Detailed reference viewed: 56 (29 ULg)A generalization of Cobham's theorem to automata over real numbers Boigelot, Bernard ; Brusten, Julien in Theoretical Computer Science (2009), 410(18), 1694-1703 Detailed reference viewed: 56 (27 ULg) |
||