P18060-N04 TTCAR - Time-Triggered Communication Architecture for an Autonomous Mobile Robot System
 
Project
Abstract
Scientists/Scholars
Project Publications
Further Activities
Cooperations
Project Results

Sensor Fusion

  • Global time allows a global interpretation of measurements

  • Time-triggered schedule supports coordinated distributed

  • Two simple algorithms for raw sensor fusion:
    • Confidence-Weighted Averaging
    • Extended Confidence Weighted Averaging
Estimated variance, fused variance

  • Robust certainty grid is a fusion algorithm combining distance measurements into a grid-like representation Obstacle detection

     

    Computer-Aided Configuration

    • In a time-triggered system all communication and computation activities have to be defined a priori

    • Idea: Liberate system designers from the burden of monotone and error prone task

    • Formalized model of components in XML

    • Formalized model of application in XML

    • Computer automated generation of schedules and code stubs

    Obstacle detection

     

    Robot Testbeds

    • TINYPHOON: Small autonomous and mobile robot in the shape of a cube with a side length of about seven centimeters

    • Different components (vision unit, decision unit, motion unit)

    • Platform serves as a testbed for generic timetriggered protocol (TTP/A)

    Tinyphoon robot
    Smartcar
  • “Smart Car” – autonomous robot with distributed control

  • 6 smart sensors

  • 5 smart actuators

  • Testbed for configuration tools

  • Data creation for sensor fusion evaluation

  • Testbed for robust certainty grid

    • Robot Soccer Simulator – official robot soccer simulation platform

    • Testbed for self-organizing multi-robot scenarios

    Robot Soccer Simulation Testbed

     

    Self-organizing Cooperating Robots

    • Self-organizing communication strategies between cooperating robots will adapt to breakdown of a robot, disruptions of the (wireless) communication system and inconsistent information
    • Self-organizing behavior cannot be designed in a straightforward way
    • Design approach:
      • Genetic evolution of desired behavior
      • Fitness function is derived from performance in simulation of target application

     

    Final Report

    W. Elmenreich: Time-Triggered Communication Architecture for an Autonomous Mobile Robot System; Public final project report, Alpen-Adria-Universität Klagenfurt, 2009

  •  
     
     

    With support from
    FWFDer Wissenschaftsfonds