Robin Pelle’s research while affiliated with University of Paris-Saclay and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (6)


Computer Aided Formal Design of Swarm Robotics Algorithms
  • Chapter

November 2021

·

25 Reads

·

2 Citations

Lecture Notes in Computer Science

Thibaut Balabonski

·

·

Robin Pelle

·

[...]

·

Previous works on formally studying mobile robotic swarms consider necessary and sufficient system hypotheses enabling to solve theoretical benchmark problems (geometric pattern formation, gathering, scattering, etc.). We argue that formal methods can also help in the early design stage of mobile robotic swarms correct-by-design protocols, even for tasks closer to real-world use cases and not previously studied theoretically. Our position is supported by a concrete case study. Starting from a real-world case scenario, we jointly design the formal problem specification, a family of protocols that are able to solve the problem, and their corresponding proof of correctness, all expressed with the same formal framework. The concrete framework we use for our development is the Pactole library based on the Coq proof assistant.


Figure 5: To avoid elimination at launch, the base waits for a robot a to be at more than D max −3D before launching another robot. On the left, as the base decides to launch a new robot, robot a is too close to robot b and will withdraw. The next round is on the right: b is now too close to c and may withdraw, leaving only c, which moves further away. As the new robot is launched, robot c should however be still visible from the base. The new robot was launched too late.
Computer Aided Formal Design of Swarm Robotics Algorithms
  • Preprint
  • File available

January 2021

·

62 Reads

Previous works on formally studying mobile robotic swarms consider necessary and sufficient system hypotheses enabling to solve theoretical benchmark problems (geometric pattern formation, gathering, scattering, etc.). We argue that formal methods can also help in the early stage of mobile robotic swarms protocol design, to obtain protocols that are correct-by-design, even for problems arising from real-world use cases, not previously studied theoretically. Our position is supported by a concrete case study. Starting from a real-world case scenario, we jointly design the formal problem specification, a family of protocols that are able to solve the problem, and their corresponding proof of correctness, all expressed with the same formal framework. The concrete framework we use for our development is the PACTOLE library based on the COQ proof assistant.

Download

Contribution à la modélisation formelle d'essaims de robots mobiles

December 2020

·

15 Reads

L’algorithmique distribuée fait partie des domaines où le raisonnement informel n’est pas une option, en particulier lorsque des erreurs dites byzantines peuvent survenir. Elle est également caractérisée par une grande diversité de modèles dont les modulations subtiles impliquent des propriétés radicalement différentes. Nous nous intéressons aux « réseaux de robots » : nuages d’entités autonomes devant accomplir une tâche en coopération. Les applications que laissent envisager ces essaims d’agents sont extrêmement prometteuses : exploration et recherche de survivants dans des zones dévastées, patrouilles et vols de drones en formation, etc. Ces quelques exemples potentiellement critiques soulignent la grande dynamicité du modèle; ils indiquent également à quel point des défaillances des robots ou des erreurs dans les protocoles distribués qui les équipent peuvent avoir de désastreuses conséquences.Pour garantir la sûreté des protocoles et la sécurité des tâches, nous visons à l’obtention, à l’aide de l’assistant à la preuve Coq, de validations mécaniques formelles de propriétés de certains protocoles distribués.Un prototype de modèle formel Coq pour les réseaux de robots, Pactole, a récemment montré la faisabilité d’une approche de vérification par assistant à la preuve dans ce cadre. Il capture assez naturellement de nombreuses variantes de ces réseaux, notamment en ce qui concerne la topologie ou les propriétés des démons. Ce modèle est bien sûr à l’ordre supérieur et s’appuie sur des types coinductifs. Il permet de démontrer en Coq à la fois des propriétés positives : le programme embarqué permet de réaliser la tâche quelle que soit la configuration de départ, comme des propriétés négatives : il n’existe aucun programme embarqué permettant de réaliser la tâche.Dans le cadre émergent des réseaux de robots, les modèles sont distingués par les caractéristiques et capacités des robots, la topologie de l’espace dans lequel ils évoluent, le degré de synchronisme (modélisé par les propriétés du démon d’activation), les erreurs pouvant survenir, etc. Le prototype Pactolen’exprime que certaines de ces variantes. Pensé dans un cadre théorique (robots ponctuels, déplacements instantanés, etc.), des hypothèses restent hors de sa portée, en particulier des hypothèses réalistes comme des exécutions totalement asynchrones ou des risques de collision. L’absence de collision est fondamentale dans toutes les applications liées aux évolutions en formation (drones) et unecondition de sécurité critique dès qu’on s’intéresse au transport aérien. Une validation formelle de cette propriété revêt donc une grande importance.Le travail consiste à étendre le modèle formel afin de prendre en compte des évolutions asynchrones de robots volumineux. Cette modélisation doit permettre une formulation aisée de protocoles et des tâches qu’ils sont censés réaliser. On s’intéressera en particulier à garantir l’absence de collision au cours de déplacements potentiellement complexes.


Continuous vs. Discrete Asynchronous Moves: A Certified Approach for Mobile Robots

September 2019

·

30 Reads

·

11 Citations

Lecture Notes in Computer Science

Oblivious Mobile Robots have been studied both in continuous Euclidean spaces, and discrete spaces (that is, graphs). However the obtained literature forms distinct sets of results for the two settings. In our view, the continuous model reflects well the physicality of robots operating in some real environment, while the discrete model reflects well the digital nature of autonomous robots, whose sensors and computing capabilities are inherently finite.We explore the possibility of bridging results between the two models. Our approach is certified using the Coq proof assistant and the Pactole framework, which we extend to the most general asynchronous model without compromising its genericity. Our extended framework is then used to formally prove the equivalence between atomic moves in a discrete space (the classical “robots on graphs” model) and non-atomic moves in a continuous unidimensional space when robot vision sensors are discrete (robots move in straigth lines between positions, but their observations are at source and destination positions only), irrespective of the problem being solved. Our effort consolidates the integration between the model, the problem specification, and its proof that is advocated by the Pactole framework.KeywordsFormal proofProof assistant Coq Mobile autonomous robotsDistributed algorithms



A Foundational Framework for Certified Impossibility Results with Mobile Robots on Graphs

January 2018

·

27 Reads

·

16 Citations

Swarms of mobile robots recently attracted the focus of the Distributed Computing community. One of the fundamental problems in this context is that of exploration: the robots must coordinate to visit all locations that are reachable from their initial positions. Despite its apparent simplicity, this problem proved quite hard to characterise fully, due to many model variants, leading to informal error-prone reasoning. Over the past few years, a significant effort permitted to set up a formal framework, relying on the Coq proof assistant, which was used to provide certified results when robots evolve in a continuous bi-dimensional Euclidean space. However, the most challenging issues with exploration arise in the discrete setting (a.k.a. graph), where locations are modeled as vertices and where edges between vertices denote the ability for a robot to move from one location to the next. We present a formal model to tackle problems and reason about robot algorithms arising in the discrete setting. Our approach extends and generalises previous research efforts focusing on the continuous model. As case studies, we consider fundamental impossibility results for exploration with stop in the discrete model. To our knowledge, those are the first certified results in this context. This framework paves the way for a general certification workflow dedicated to mobile robots on graphs.

Citations (3)


... Although the seminal paper [32] focused on continuous spaces, many recent papers [19] consider robots evolving on a discrete graph (that is, robots are located on a discrete set of locations, the nodes of the graph, and may move from one location to the next if an edge exists between the two locations), as it was recently observed that discrete observations model better actual sensing devices [2]. For the particular topology we consider, the grid, many problems were previously investigated, e.g., exploration [14,15], perpetual exploration [4], scattering [3], dispersion [28], gathering [11], mutual visibility [1], pattern formation [5], and convex hull formation [21]. ...

Reference:

An Asynchronous Maximum Independent Set Algorithm by Myopic Luminous Robots on Grids
Continuous vs. Discrete Asynchronous Moves: A Certified Approach for Mobile Robots
  • Citing Chapter
  • September 2019

Lecture Notes in Computer Science

... When robots move freely in a continuous two-dimensional Euclidean space (as considered in this paper), to the best of our knowledge the only formal framework available is Pactole. 2 It relies on higherorder logic to certify impossibility results [3,12,5], as well as the correctness of algorithms [13,16] in the FSYNC and SSYNC models, possibly for an arbitrary number of robots (hence in a scalable manner). Pactole was recently extended by Balabonski et al. [4] to handle the ASYNC model, thanks to its modular design. However, in its current form, Pactole lacks automation; that is, in order to prove a result formally, one still has to write the proof (that is automatically verified), which requires expertise both in Coq (the language Pactole is based upon) and about the mathematical and logical arguments one should use to complete the proof. ...

Brief Announcement Continuous vs. Discrete Asynchronous Moves: A Certified Approach for Mobile Robots: 20th International Symposium, SSS 2018, Tokyo, Japan, November 4–7, 2018, Proceedings
  • Citing Chapter
  • October 2018

Lecture Notes in Computer Science

... Formal methods encompass a long-lasting path of research that is meant to overcome errors of human origin. Unsurprisingly, this mechanized approach to protocol correctness was used in the context of mobile robots [3,4,5,6,7,8,9,14,19,20,37,38,39,40]. ...

A Foundational Framework for Certified Impossibility Results with Mobile Robots on Graphs
  • Citing Conference Paper
  • January 2018