Data from Pioneer Array

Transcription

1 Data from Pioneer Array Arjuna Balasuriya Dept. of Mechanical Engineering, MIT

2 Contents Ocean Observatory Initiatives (OOI) CI: Planning and Prosecution Sub-system On-board Autonomy: MOOS-IvP & ASPEN/CASPER Conclusions Arjuna Balasuriya Page 2

3 Ocean Observatory Initiative - NSF Arjuna Balasuriya Page 3

4 Pioneer Array 3 Surface Moorings 2 with Docks for Autonomous Underwater Vehicles (AUVs) 7 Profiler Moorings 6 Gliders 3 AUVs Communication 100 m 40 km Inductive along Mooring cables Connector for AUVs SatComm from Gliders, Surface Moorings, and Profiler Moorings Acomms 15 km 500 Arjuna Balasuriya Page 4

5 CI Subsystems User Application Subsystems Infrastructure Subsystems Arjuna Balasuriya Page 5

6 CI Sub-Systems Common Operating Infrastructure Identity Management Governance Framework State Management Resource Management Service Framework Presentation Framework Exchange Common Execution Infrastructure Data Management (Information Distribution) Data Management (Science) Sensing & Acquisition Analysis & Synthesis Planning & Prosecution Arjuna Balasuriya Page 6

7 CI: Planning & Prosecution Sub System Arjuna Balasuriya Page 7

8 Autonomous Control Arjuna Balasuriya Page 8

9 Adaptive and Collaborative Ocean Observation Observatory Management Feature Tracks Feature Tracks 80bps CCL Cluster activated by Shore Station Environmental updates - ASPEN AUVs & gliders approach the ROI. Adaptively sense and exploit environment. Detect features and adaptively manoeuvre to resolve ambiguity for optimal tracking and localization (DLT) Collaborative Localization and Tracking (LT). Cluster formations Transmit updated, enhanced measurement messages. Environmental updates in real-time. Arjuna Balasuriya Page 9

10 Vehicle Autonomy Architecture Backseat Driver Paradigm - ASTM F41 Three components of the overall vehicle architecture. Control and Navigation (frontseat driver) Actuator control, inertial navigation, GPS, compass, DVL, dead-reckoning systems, vehicle safety. Autonomy System as a Whole Sensor processing, sensor fusion, autonomy, contact management, data logging, system monitoring, mission control, communication. Autonomous Decision-Making (backseat driver) Deciding vehicle heading, speed, and depth. Autonomy System as a Whole MOOS Autonomous Decision- Making IvP Helm Control and Navigation System Payload Computer Main Vehicle Computer MOOS-IvP Backseat Driver Vehicles BF21 (with towed VSA,HLA) IVAR Ocean Server (w/ HLA) SCOUT (MIT kayaks) REMUS 100/600 (In progress, PLUS, UCCI) OEX (NURC, w SLITA HLA) FOLEGA (NURC, Univ. Pisa) MOOS-IvP Mandated Programs UCCI (ONR 07-11) Distributed MCM PLUS (ONR 08-12) Undersea Surveillance, ASW, OOI (NSF 08-12) Ocean Observatory Initiative Arjuna Balasuriya Page 10

11 Vehicle Autonomy Architecture Three components of the overall vehicle architecture. Control and Navigation (frontseat driver) Actuator control, inertial navigation, GPS, compass, DVL, dead-reckoning systems, vehicle safety. Autonomy System as a Whole Sensor processing, sensor fusion, autonomy, contact management, data logging, system monitoring, mission control, communication. Autonomous Decision-Making (backseat driver) Deciding vehicle heading, speed, and depth. Autonomy System as a Whole MOOS Autonomous Decision- Making IvP Helm Control and Navigation System Payload Computer Main Vehicle Computer module module module MOOS The glue for the autonomy system as a whole. module module module MOOS Core module module module module Modules coordinated through a publish and subscribe interface. Overall system is built incrementally. module Publish Subscribe MOOSDB Arjuna Balasuriya Page 11

12 On-Board Autonomy Three components of the overall vehicle architecture. Control and Navigation (frontseat driver) Actuator control, inertial navigation, GPS, compass, DVL, dead-reckoning systems, vehicle safety. Autonomy System as a Whole Sensor processing, sensor fusion, autonomy, contact management, data logging, system monitoring, mission control, communication. Autonomous Decision-Making (backseat driver) Deciding vehicle heading, speed, and depth. behavior behavior behavior behavior behavior IvP Core behavior behavior behavior behavior behavior Autonomy System as a Whole MOOS Autonomous Decision- Making IvP Helm Control and Navigation System IvP Helm Payload Computer Main Vehicle Computer The glue for the autonomous decision-making engine Modules coordinated through logic (behavior algebra), objective functions and multi-objective optimization. Overall system is built incrementally. behavior Sensor Info Objective function IvP Solver Arjuna Balasuriya Page 12

13 Observing System Simulation Experiment (OSSE) Arjuna Balasuriya Page 13

14 Observing System Simulation Experiment (OSSE) OSSE Management ASPEN Glider Operations Control & Data plans Casper plans AUV Operations Control & Data ROMS data Control & Data exchange Iridium phone bank data satellites CODAR Non-MOOS gliders Plans, controls, and Data Glider AUV Casper MOOS-IvP Platform Controller Casper MOOS-IvP Platform Controller Arjuna Balasuriya Page 14

15 On-board Autonomy Space segment also acquires data and can be tasked in response Onboard Mission Planning and Reactive Control (on each node) MOOS Plans or commands to control SW Observatory Facility Services (shore side planner) ASPEN Autonomous Coordinated Response via Shore Planner Responses Event Response Services Multi Objective Observation Catalog On-board Autonomy with CASPER Data & event notices Event DB/Model Arjuna Balasuriya Page 15

16 Autonomy OSSE 09 Arjuna Balasuriya Page 16

17 AUV Operations Setup for OSSE09 MIT JPL Google Earth 2.4 Mbps MIT Top Side MOOSDB 2.4GHz R/V Arabella Gateway Buoy Greatwhite IVER 2 80 bps Hammerhead IVER 1 80 bps 80 bps REMUS 80 bps Arjuna Balasuriya Page 17

18 AUVs in Operation Greatwhite Cal. Poly Remus Rutgers Hammerhead MIT/NUWC Op-Box determined by ASPEN MOOS-IvP autonomy on-board each IVER AUV Real-Time communication between AUVs and shore using acoustic modems Environmental sampling using CTD Arabella track Hammerhead track Arjuna Balasuriya Page 18

19 Behavior Implementation Gliders vs. (actively propelled) AUVs Currently have a library of AUV behaviors (basic & advanced) in MOOS IvP Yoyo Waypoint Tracking etc. Approach Created a glider behavior interface for MOOS IvP Use existing AUV behaviors Arjuna Balasuriya Page 19

20 Eg 1: Thermocline Tracking An adaptive environmental sampling and tracking behavior employing REA Calculates and monitors the changing temperature gradient through the water column based on the depth vs. temperature profile Vehicle autonomously adapts the depth range of its yo-yo to more closely sample and track a thermocline Successfully executed during the GLINT 09 experiment Arjuna Balasuriya Page 20

21 In Situ CTD Gradient Determination Arjuna Balasuriya Page 21

22 Eg. 2: Front Tracking Arjuna Balasuriya Page 22

23 Conclusions Construction of the Ocean Observatory Initiative Mobile platforms (robots) Adaptive Ocean sampling React to events Collaborative/cooperative net-centric autonomy Nested autonomy backseat driver paradigm net-centric autonomy with challenging comms. Infrastructure Arjuna Balasuriya Page 23

24 Thank You Arjuna Balasuriya Page 24