Special Interest Group on NETworking SIGNET Discovery, localization, and recognition of smart objects by a mobile robot UNIVERSITY OF PADUA Dept. of Information.

Slides:



Advertisements
Similar presentations
Special Interest Group on NETworking SIGNET Range-only SLAM with a Mobile Robot and a Wireless Sensor Networks UNIVERSITY OF PADUA Dept. of information.
Advertisements

Giovanni Zanca, Francesco Zorzi, Andrea Zanella and Michele Zorzi
Localization with RSSI Method at Wireless Sensor Networks Osman Ceylan Electronics Engineering PhD Student, Istanbul Technical University, Turkiye
Decentralized Reactive Clustering in Sensor Networks Yingyue Xu April 26, 2015.
Wearable Badge for Indoor Location Estimation of Mobile Users MAS 961 Developing Applications for Sensor Networks Daniel Olguin Olguin MIT Media Lab.
The WolfBot: Swarm Robotics Platform with Image Processing Applications Jimit Patel Dr. Edgar Lobaton Department of Electrical and Computer Engineering.
Discussion topics SLAM overview Range and Odometry data Landmarks
Computer Networks Group Universität Paderborn Ad hoc and Sensor Networks Chapter 9: Localization & positioning Holger Karl.
Kiyoshi Irie, Tomoaki Yoshida, and Masahiro Tomono 2011 IEEE International Conference on Robotics and Automation Shanghai International Conference Center.
Real-Time Video Analysis on an Embedded Smart Camera for Traffic Surveillance Presenter: Yu-Wei Fan.
Visual Navigation in Modified Environments From Biology to SLAM Sotirios Ch. Diamantas and Richard Crowder.
1 Introduction to Wireless Sensor Networks. 2 Learning Objectives Understand the basics of Wireless Sensor Networks (WSNs) –Applications –Constraints.
An Empirical Characterization of Radio Signal Strength Variability in 3-D IEEE Networks Using Monopole Antennas Dimitrios Lymberopoulos, Quentin.
1 University of Freiburg Computer Networks and Telematics Prof. Christian Schindelhauer Wireless Sensor Networks 16th Lecture Christian Schindelhauer.
Probabilistic Robotics
An robotic sculpture speaking to people An robotic sculpture speaking to people “Would you like a drink?” Fiberglass and robotics Albano Guatti, 2005 University.
Part 2 of 3: Bayesian Network and Dynamic Bayesian Network.
Generic Sensor Platform for Networked Sensors Haywood Ho.
Enhancing RSSI-based Tracking Accuracy in Wireless Sensor Networks
8/22/20061 Maintaining a Linked Network Chain Utilizing Decentralized Mobility Control AIAA GNC Conference & Exhibit Aug. 21, 2006 Cory Dixon and Eric.
1 Localization Technologies for Sensor Networks Craig Gotsman, Technion/Harvard Collaboration with: Yehuda Koren, AT&T Labs.
Authors: Joseph Djugash, Sanjiv Singh, George Kantor and Wei Zhang
CC2420 Channel and RSSI Evaluation Nov/22/2006 Dept. of EECS, UC Berkeley C O nnect vityLab i.
Simultaneous Localization and Map Building System for Prototype Mars Rover CECS 398 Capstone Design I October 24, 2001.
1 Range-Only SLAM for Robots Operating Cooperatively with Sensor Networks Joseph Djugash Sanjiv Singh George Kantor Wei Zhang Carnegie Mellon University.
California Car License Plate Recognition System ZhengHui Hu Advisor: Dr. Kang.
Autonomous discovery, localization and recognition of smart objects through WSN and image features E. Menegatti, M. Mina, A. Pretto P. Zanuttigh, S. Zanconato.
SensEye: A Multi-Tier Camera Sensor Network by Purushottam Kulkarni, Deepak Ganesan, Prashant Shenoy, and Qifeng Lu Presenters: Yen-Chia Chen and Ivan.
Fuzzy control of a mobile robot Implementation using a MATLAB-based rapid prototyping system.
Kalman filter and SLAM problem
Zereik E., Biggio A., Merlo A. and Casalino G. EUCASS 2011 – 4-8 July, St. Petersburg, Russia.
Robot Compagnion Localization at home and in the office Arnoud Visser, Jürgen Sturm, Frans Groen University of Amsterdam Informatics Institute.
Introduction to Sensor Networks Rabie A. Ramadan, PhD Cairo University 3.
 Copyright 2006 Digital Enterprise Research Institute. All rights reserved. Large-Scale Sensor Networks Prof. Manfred Hauswirth (joint work.
Tour Guide Robot Project Face Detection and Face Orientation on The Mobile Robot Robotino Gökhan Remzi Yavuz Ayşenur Bilgin.
3D SLAM for Omni-directional Camera
WMNL Sensors Deployment Enhancement by a Mobile Robot in Wireless Sensor Networks Ridha Soua, Leila Saidane, Pascale Minet 2010 IEEE Ninth International.
Tracking with Unreliable Node Sequences Ziguo Zhong, Ting Zhu, Dan Wang and Tian He Computer Science and Engineering, University of Minnesota Infocom 2009.
Mobile Relay Configuration in Data-Intensive Wireless Sensor Networks.
Introduction to Wireless Sensor Networks
Robust localization algorithms for an autonomous campus tour guide Richard Thrapp Christian Westbrook Devika Subramanian Rice University Presented at ICRA.
Young Ki Baik, Computer Vision Lab.
 Copyright 2006 Digital Enterprise Research Institute. All rights reserved. Global Sensor Networks Manfred Hauswirth (joint work with Karl.
Link layer Murat Demirbas SUNY Buffalo CSE Dept..
Self–localization of Wireless Sensor Nodes by means of Autonomous Mobile Robots A note on the use of these ppt slides: We’re making these slides freely.
Visual SLAM Visual SLAM SPL Seminar (Fri) Young Ki Baik Computer Vision Lab.
College of Engineering Anchor Nodes Placement for Effective Passive Localization Karthikeyan Pasupathy Major Advisor: Dr. Robert Akl Department of Computer.
Wireless Sensor Network (WSN). WSN - Basic Concept WSN is a wireless network consisting of spatially distributed autonomous devices using sensors to cooperatively.
1 Research Question  Can a vision-based mobile robot  with limited computation and memory,  and rapidly varying camera positions,  operate autonomously.
Human Tracking System Using DFP in Wireless Environment 3 rd - Review Batch-09 Project Guide Project Members Mrs.G.Sharmila V.Karunya ( ) AP/CSE.
Real-Time Simultaneous Localization and Mapping with a Single Camera (Mono SLAM) Young Ki Baik Computer Vision Lab. Seoul National University.
Positioning in Ad-Hoc Networks - A Problem Statement Jan Beutel Computer Engineering and Networks Lab Swiss Federal Institute of Technology (ETH) Zurich.
1 A Context Discovery Middleware for Context-Aware Applications with Heterogeneous Sensors Yu-Min Tseng.
Abstract A Structured Approach for Modular Design: A Plug and Play Middleware for Sensory Modules, Actuation Platforms, Task Descriptions and Implementations.
Range-Only SLAM for Robots Operating Cooperatively with Sensor Networks Authors: Joseph Djugash, Sanjiv Singh, George Kantor and Wei Zhang Reading Assignment.
Syed Hassan Ahmed Syed Hassan Ahmed, Safdar H. Bouk, Nadeem Javaid, and Iwao Sasase RIU Islamabad. IMNIC’12, RIU Islamabad.
1/24 Experimental Analysis of Area Localization Scheme for Sensor Networks Vijay Chandrasekhar 1, Zhi Ang Eu 1, Winston K.G. Seah 1,2 and Arumugam Pillai.
Thermal Detecting Wireless Sensor Network
Towards the autonomous navigation of intelligent robots for risky interventions Janusz Bedkowski, Grzegorz Kowalski, Zbigniew Borkowicz, Andrzej Masłowski.
Combined Human, Antenna Orientation in Elevation Direction and Ground Effect on RSSI in Wireless Sensor Networks Syed Hassan Ahmed, Safdar H. Bouk, Nadeem.
Introduction to Wireless Sensor Networks
Software Architecture of Sensors. Hardware - Sensor Nodes Sensing: sensor --a transducer that converts a physical, chemical, or biological parameter into.
CSE-473 Mobile Robot Mapping.
Simultaneous Localization and Mapping
Fast Localization for Emergency Monitoring and Rescue in Disaster Scenarios Based on WSN SPEAKER:Jyun-Ying Yu ADVISOR:DR. Kai-Wei Ke DATE:2018/05/04.
Wireless Mesh Networks
Developing different types of Motes in WSN
Overview: Chapter 4 Infrastructure Establishment
Probabilistic Robotics Bayes Filter Implementations FastSLAM
Task Manager & Profile Interface
Presentation transcript:

Special Interest Group on NETworking SIGNET Discovery, localization, and recognition of smart objects by a mobile robot UNIVERSITY OF PADUA Dept. of Information Engineering E. Menegatti M. Danieletto, M. Mina, A. Pretto, A. Bardella, A. Zanella, P. Zanuttigh Intelligent Autonomous Systems Lab University of Padua

Recognition of smart objects Goal: A robotic system exploiting Wireless Sensor Network (WSN) technologies for implementing an ambient intelligence scenario. We address the problems of object discovery, localization, and recognition in a fully distributed way. The robot does not have any a priori information! Neither on the number nor on the kind of objects in the environment.

3 RAMSES2 - Project RAMSES 2 : integRation of Autonomous Mobile robots and wireless SEnsor networks for Surveillance and reScue A utonomous M obile R obot Wireless network channel W ireless S ensors N etwork Laptop eyesIFX motes from Infineon b wireless channel

4 Experimental Set up EyesIFX sensor nodes –Infineon Technologies. –19.2 kbps bit 868 MHz –Light, temperature, RSSI sensors AMR Bender – self-made, based on Pioneer 2 ActivMedia platform – Linux OS with Miro middleware – ATX motherboard – 1,6 GHz Intel Pentium 4, 256 MB RAM, 160 GB HD EyesIFX connected to ATX via USB + EyesService class added to Miro – Omnidirectional camera, odometers

5 Step 1 - DISCOVERY Allow a bidirectional serial communication (ASCII chars) Allow robot’s applications to interact with the WSN  Physical connection between robot and mote  Serial port emulation over USB (VCP)  Standard commands for eyesIFX sensor  Predefined actions to access to the WSN Input/OutputFunctions

6 Middleware Miro The robot is programmed exploiting the framework Miro Miro is a framework for mobile robot programming developed by Gerd Mayer and Gerhard Kraetzschmar at Ulm University Miro is a middleware based on CORBA architecture for creating and managing distributed services. Miro is based on TAO libraries of the ACE framework. We interact with the eyesIFX mote on board of the robot through a Miro service we created, called EyesService. Range-only SLAM with a Mobile Robot and a Wireless Sensor Networks 6

Why Localization? WSN deploying is an annoying and time consuming task. Moreover, motes can be attached to objects that are moved around Step 2 - LOCALIZATION First goal of the project localize WSN nodes spread in unknown positions inside a building using a mobile robot.

Localization Approaches Three main ranging approaches: Angle of Arrival Time of Arrival Received Signal Strength Indicator (RSSI) Focus on RSSI: No specific Hardware required Poor outdoor ranging performance  Very poor indoor ranging performance 

Our Solution SLAM (Simultaneous Localization And Mapping), for a mobile robot moving in an unknown environment in which there is a WSN (Wireless Sensor Network). We use only: robot’s odometry; range measurements from the nodes to the robot Step 2 - LOCALIZATION

For the same range, we can measure very different RSSIWe measure the RSSI to estimate the range, then... RSSI based ranging is VERY noisy! Random variations due to shadowing and fading obscure the log-decreasing law for the received power vs distance How harsh is the indoor radio channel?

Experiments EyesIFX v2 Mote Robot “Bender” 10 x 6 m environment

Results (1/4) - SLAM Much better that classical static WSN localization algorithm Large variance on residual error for motes locations Slightly better results taking only highest RSSI measurements (Elab 2) Fig. 1 residual mean error on robot and motes position

Much better that classical static WSN localization algorithm Large variance on residual error for motes locations Results (2/4) - SLAM

If we correctly initialize the mote position in the EKF...( Elab 5 & 6 ) Results: Slight improvements on robot residual error Large improvements on mote residual error Fig. 2 Residual mean error on robot and motes position Where does the error come from?

Results (4/4) - SLAM

16 Extended Kalman Filter OdometryRSSI Measures Initialization Particle Filter Motes pose and robot position estimation Localization with delayed particle filter Localization Algorithm:

Delayed Initialization based on Particle Filter 15

18 Step 3 - RECOGNITION Goal: the robot should be able to recognize the objects from a description of the objects’ appearance. The object appearance is stored inside the object (in the motes). The appearances is coded by scale invariant feature descriptors robust also to motion blur (MoBIF) Feature matching under motion blur (Pretto et al. ICRA 2009)

19 MoBIF Matching The object in the robot’s camera image are identified matching the MoBIF descriptors (robust to motion blur) transmitted by the smart objects with the MoBIF descriptor extracted from the robot image; A minimum number of matching is needed to correctly identify an object; matching is robust to scale, occlusion, illumination and rotation;

20 Storing the appearance To achieve robustness to scale: the object is imaged at distances: near, medium, far and the MoBIF descriptor are merged To achieve robustness to rotation: the object is imaged every 20 deg. and the MoBIF descriptor are merged The merging of the MoBIF removes redundancy! 5 meters 2 meters 1 meter

21 The robot uses a simple visual-servoing image map to approach the object. An approaching direction is identified for each of the 9 regions of the image (the red arrow). Future work: IF the robot camera is calibrated the object location in the image can be feed to the SLAM algorithm to further refine the object pose estimation. Step 4 - APPROACHING

Special Interest Group on NETworking SIGNET Discovery, localization, and recognition of smart objects by a mobile robot UNIVERSITY OF PADUA Dept. of Information Engineering E. Menegatti M. Danieletto, M. Mina, A. Pretto, A. Bardella, A. Zanella, P. Zanuttigh Intelligent Autonomous Systems Lab University of Padua