A Randomized Gathering Algorithm for Multiple Robots with Limited Sensing Capabilities Noam Gordon Israel A. Wagner Alfred M. Bruckstein Technion – Israel.

Slides:



Advertisements
Similar presentations
Friedhelm Meyer auf der Heide 1 HEINZ NIXDORF INSTITUTE University of Paderborn Algorithms and Complexity Local Strategies for Building Geometric Formations.
Advertisements

Friedhelm Meyer auf der Heide 1 HEINZ NIXDORF INSTITUTE University of Paderborn Algorithms and Complexity Energy-Efficient Local Strategies for Robotic.
Routing Complexity of Faulty Networks Omer Angel Itai Benjamini Eran Ofek Udi Wieder The Weizmann Institute of Science.
Completeness and Expressiveness
Solving connectivity problems parameterized by treewidth in single exponential time Marek Cygan, Marcin Pilipczuk, Michal Pilipczuk Jesper Nederlof, Dagstuhl.
The Capacity of Wireless Networks Danss Course, Sunday, 23/11/03.
Shortest Vector In A Lattice is NP-Hard to approximate
Distributed Computing 1. Lower bound for leader election on a complete graph Shmuel Zaks ©
1 Lecture 5 (part 2) Graphs II Euler and Hamiltonian Path / Circuit Reading: Epp Chp 11.2, 11.3.
A Topological Approach to Voxelization Samuli Laine NVIDIA.
Self Stabilizing Algorithms for Topology Management Presentation: Deniz Çokuslu.
Computational Geometry, WS 2007/08 Lecture 17 Prof. Dr. Thomas Ottmann Algorithmen & Datenstrukturen, Institut für Informatik Fakultät für Angewandte Wissenschaften.
Beyond Trilateration: On the Localizability of Wireless Ad Hoc Networks Reported by: 莫斌.
I. The Problem of Molding Does a given object have a mold from which it can be removed? object not removable mold 1 object removable Assumptions The object.
Convex Hulls in Two Dimensions Definitions Basic algorithms Gift Wrapping (algorithm of Jarvis ) Graham scan Divide and conquer Convex Hull for line intersections.
CPSC 689: Discrete Algorithms for Mobile and Wireless Systems Spring 2009 Prof. Jennifer Welch.
Tirgul 8 Graph algorithms: Strongly connected components.
A (1+  )-Approximation Algorithm for 2-Line-Center P.K. Agarwal, C.M. Procopiuc, K.R. Varadarajan Computational Geometry 2003.
Gathering Multiple Robotic A(ge)nts with limited visibility Noam Gordon Israel A. Wagner Alfred M. Bruckstein Technion - IIT.
Integration in the Complex Plane CHAPTER 18. Ch18_2 Contents  18.1 Contour Integrals 18.1 Contour Integrals  18.2 Cauchy-Goursat Theorem 18.2 Cauchy-Goursat.
A Randomized Polynomial-Time Simplex Algorithm for Linear Programming Daniel A. Spielman, Yale Joint work with Jonathan Kelner, M.I.T.
Discrete geometry Lecture 2 1 © Alexander & Michael Bronstein
Data Transmission and Base Station Placement for Optimizing Network Lifetime. E. Arkin, V. Polishchuk, A. Efrat, S. Ramasubramanian,V. PolishchukA. EfratS.
Asymmetric Ramsey Properties of Random Graphs involving Cliques Reto Spöhel Joint work with Martin Marciniszyn, Jozef Skokan, and Angelika Steger TexPoint.
Jie Gao Joint work with Amitabh Basu*, Joseph Mitchell, Girishkumar Stony Brook Distributed Localization using Noisy Distance and Angle Information.
1 Motion Planning Algorithms : BUG-family. 2 To plan a path  find a continuous trajectory leading from initial position of the automaton (a mobile robot)
Clustering Color/Intensity
Complexity 19-1 Complexity Andrei Bulatov More Probabilistic Algorithms.
A Hierarchical Energy-Efficient Framework for Data Aggregation in Wireless Sensor Networks IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 55, NO. 3, MAY.
MCA 520: Graph Theory Instructor Neelima Gupta
Geometric characterization of nodal domains Y. Elon, C. Joas, S. Gnutzman and U. Smilansky Non-regular surfaces and random wave ensembles General scope.
UMass Lowell Computer Science Advanced Algorithms Computational Geometry Prof. Karen Daniels Spring, 2007 O’Rourke Chapter 8 Motion Planning.
Geometric Probing with Light Beacons on Multiple Mobile Robots Sarah Bergbreiter CS287 Project Presentation May 1, 2002.
Comment reconstruire le graphe de visibilité d’un polygone? (Reconstructing visibility graphs of polygons) Jérémie Chalopin, Shantanu Das LIF, Aix-Marseille.
Graph Theory Chapter 6 Planar Graphs Ch. 6. Planar Graphs.
On Probabilistic Snap-Stabilization Karine Altisen Stéphane Devismes University of Grenoble.
Discrete Bee Dance Algorithms for Pattern Formation on the Grid Noam Gordon Israel A. Wagner Alfred M. Bruckstein Technion IIT, Israel.
On Probabilistic Snap-Stabilization Karine Altisen Stéphane Devismes University of Grenoble.
DATA MINING LECTURE 13 Absorbing Random walks Coverage.
Ran El-Yaniv and Dmitry Pechyony Technion – Israel Institute of Technology, Haifa, Israel Transductive Rademacher Complexity and its Applications.
Franck Petit INRIA, LIP Lab. Univ. / ENS of Lyon France Optimal Probabilistic Ring Exploration by Semi-Synchronous Oblivious Robots Joint work with Stéphane.
Edge-disjoint induced subgraphs with given minimum degree Raphael Yuster 2012.
1 Maximal Independent Set. 2 Independent Set (IS): In a graph G=(V,E), |V|=n, |E|=m, any set of nodes that are not adjacent.
Salah A. Aly,Moustafa Youssef, Hager S. Darwish,Mahmoud Zidan Distributed Flooding-based Storage Algorithms for Large-Scale Wireless Sensor Networks Communications,
The Fundamental Theorem for Line Integrals. Questions:  How do you determine if a vector field is conservative?  What special properties do conservative.
An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks Seema Bandyopadhyay and Edward J. Coyle Presented by Yu Wang.
By J. Burns and J. Pachl Based on a presentation by Irina Shapira and Julia Mosin Uniform Self-Stabilization 1 P0P0 P1P1 P2P2 P3P3 P4P4 P5P5.
Markov Chains and Random Walks. Def: A stochastic process X={X(t),t ∈ T} is a collection of random variables. If T is a countable set, say T={0,1,2, …
1 Leader Election in Rings. 2 A Ring Network Sense of direction left right.
Chapter 11 Resource Allocation by Mikhail Nesterenko “Distributed Algorithms” by Nancy A. Lynch.
Unique Games Approximation Amit Weinstein Complexity Seminar, Fall 2006 Based on: “Near Optimal Algorithms for Unique Games" by M. Charikar, K. Makarychev,
Complexity and Efficient Algorithms Group / Department of Computer Science Testing the Cluster Structure of Graphs Christian Sohler joint work with Artur.
1 Distributed Vertex Coloring. 2 Vertex Coloring: each vertex is assigned a color.
Introduction Wireless Ad-Hoc Network  Set of transceivers communicating by radio.
Theory of Computational Complexity Probability and Computing Lee Minseon Iwama and Ito lab M1 1.
1 Lecture 5 (part 2) Graphs II (a) Circuits; (b) Representation Reading: Epp Chp 11.2, 11.3
Polygon Triangulation
Theory of Computational Complexity Probability and Computing Ryosuke Sasanuma Iwama and Ito lab M1.
Computational Geometry
Beth Tsai Jennifer E. Walter Nancy M. Amato Department of Computer Science Texas A&M University, College Station Distributed Reconfiguration of Metamorphic.
Theory of Computational Complexity M1 Takao Inoshita Iwama & Ito Lab Graduate School of Informatics, Kyoto University.
Krishnendu ChatterjeeFormal Methods Class1 MARKOV CHAINS.
Theory of Computational Complexity Probability and Computing Chapter Hikaru Inada Iwama and Ito lab M1.
4/22/20031/28. 4/22/20031/28 Presentation Outline  Multiple Agents – An Introduction  How to build an ant robot  Self-Organization of Multiple Agents.
3. Polygon Triangulation
Markov Chains and Random Walks
Depth Estimation via Sampling
I. The Problem of Molding
Introduction Wireless Ad-Hoc Network
Locality In Distributed Graph Algorithms
Presentation transcript:

A Randomized Gathering Algorithm for Multiple Robots with Limited Sensing Capabilities Noam Gordon Israel A. Wagner Alfred M. Bruckstein Technion – Israel Institute of Technology

The Gathering Problem How to make multiple autonomous robots gather in a small region/point? a.k.a. Point Formation or Convergence. Fundamental to formation and self-organization problems. Useful for collecting robots after a mission or after being initially dispersed. Useful for nano-robot aggregation.

Current Works Suzuki et al. ’96–’99 Prencipe et al. ’01–’03 Bruckstein et al. ’91–’03 Francis et al. ’03–’04

The World Model The agents are  points in the plane.  “semi-synchronous” – moving in synchronous steps, but randomly scheduled to act only during some steps.  anonymous, homogeneous, memoryless.  able to move up to a distance σ in one step. An agent can see only up to a distance V. An agent cannot measure the distance, but rather only the direction toward a nearby agent.

Maintaining Visibility abV

a V

a

The intersection of circles is empty. The agent is “surrounded” and cannot move. If there are no visible agents, then the allowable region is a full disc. a

A deterministic algorithm In previous work (‘04), we proposed a deterministic algorithm. Related to an algorithm by Sugihara et al (‘96). a

The Proposed Randomized Algorithm a

The agents simply move randomly while maintaining visibility. Similar global behavior remains.  Phase 1: Non-uniform contraction of the swarm, assuming a rough large-scale polygonal shape, ultimately becoming a small dense cluster.  Phase 2: The cluster stops contracting and begins wandering in the plane.

Proof of convergence Lemma 1: If only a 1 is active and it moves a step s closer to COM, then the variance ∑d i 2 will decrease by at least |s| 2 /n. a1a1 d1d1 s θ

Proof of convergence d = (d 1,…,d n ) T = displacement from COM d T d is the variance. s = (s,0,…,0) T = agents’ steps (only a 1 moves) d’ = displacement from COM after move. 1 = (1,…,1) T d’ = d – s + 1·s/n

Proof of convergence d’ = d – s + 1·s/n d’ T d’ = (d – s + 1·s/n) T (d – s + 1·s/n) = d T d + s(s-2d) – s T s/n = d T d + |s|(2|d 1 |cosθ-|s|)-|s| 2 /n a 1 moves closer to COM iff the second term is negative. In this case, d’ T d’ - d’ T d’ ≤ -|s| 2 /n

Proof of convergence Lemma 2: There exists a vertex a i in CH with bounded angle φ i ≤ φ* 0 from the center of mass. Denote w.l.o.g. a 1 = a i. aiai d i ≥ d* > 0 φ i ≤ φ* < π

Proof of convergence Lemma 3: There exist p*,s* > 0, such that a 1 will choose to move a step of length > s*, closer to COM, with probability > p*. a1a1 COM ∂CH

Proof of convergence Theorem: The agents will gather and remain forever within a cluster of diameter < V, given a connected initial visibility graph. Proof: From above lemmas, there always exists an agent a 1, such that if it is the only active agent, then the variance will decrease by at least s* 2 /n, with probability p*. From our strong asynchronicity assumption, the probability that indeed a 1 will be the only active agent, is at least some ε > 0.

Proof of convergence Thus, there is always a probability > εp* that the variance will decrease by at least s* 2 /n. The variance is bounded from above. Therefore, it will decrease arbitrarily with probability 1 and within finite expected time. The diameter will decrease along with the variance. Once it reaches V (=full visibility), it will remain bounded by V, since the algorithm maintains visibility.

The Wandering Cluster Once the swarm has contracted into a small dense cluster, it begins to wander in the plane. The cluster is so small, that the outer agents (at the corners of CH) most often leap over the cluster instead of moving inside it. The random scheduling and choices make the cluster wander randomly.

The Wandering Cluster Is the cluster’s random walk recurrent? We need to check whether the resulting random walk is biased or not. n=1: A solitary agent performs a uniform random walk by definition. n=2: Due to symmetry, the relative orientation of two agents is itself recurrent. Hence, initial orientation is forgotten and the random walk becomes unbiased. n≥3: We conjecture that the cluster’s random walk is indeed recurrent.

The Wandering Cluster If our conjecture is true, then the algorithm will work for all initial conditions! Each connected component of the initial visibility graph will contract to an independent cluster. All clusters will eventually merge, due to their recurrence.

Conclusions A simple yet powerful randomized gathering algorithm for myopic robots which cannot measure distances. A vivid example of how very simple local behaviors turn into complex global behaviors. Interestingly, the agents just move randomly while keeping visibility. Maintain visibility → Gather So is this the only thing that these humble agents can do?