Team Kanaloa: research initiatives and the Vertically Integrated Project (VIP) development paradigm

Transcription

1 Additive Manufacturing Renewable Energy and Energy Storage Astronomical Instruments and Precision Engineering Team Kanaloa: research initiatives and the Vertically Integrated Project (VIP) development paradigm Brennan Yamamoto, Allison Wong rip.eng.hawaii.edu Industrial Automation and Machine Design

2 Vertically Integrated Projects A long-term research and development paradigm started at the Georgia Institute of Technology (14 current member universities). Purpose: extend the academic research and design process beyond the typical semester/year-long course, by engaging students in multi-year research projects. Desired outcomes: Development of field-specific, core research competencies earlier in the education process. Wider impact of invested research dollars (more explicit trickle down effect). Increased capability to handle wide project scope (multi-discipline), and logistically demanding research topics.

3 Team Kanaloa: Unmanned X Systems Vertically Integrated Project dedicated to development of unmanned systems technologies (aerial/surface/underwater). Purpose: to advance novel technologies in unmanned systems (interoperability, sensing, navigation, control and autonomy) and develop open-source unmanned systems as modular, mission adaptable vehicles for industry and research use. Fall 2016 to present: 110 participating students in biology, biological E, civil E, computer E, computer science, electrical E, ocean resource E, mechanical E, and physics...

4 Kanaloa Program Architecture Mechanical design Technical competency Program Manager Project Managers, Subsystem Integrators, Finance Manager subsystems Hardware Selection and Interoperability Signal Processing and Sensor Fusion Guidance and Control Mission Planning Subsystem & Technical Leads Subsystem Members

5 Unmanned Systems Core Challenges The study and application of mobile robotics and unmanned systems can be categorized as follows: Increasing integration complexity Hardware Interoperability How shall I think and communicate? Sensor and Actuator Selection How shall I perceive and interact with the environment? Sensor Fusion and Signal Processing How shall I interpret noisy & incomplete sensor data? Guidance, Control and Path Planning How shall I command my motion? Mission Planning and Autonomy How shall I decide what to do next? Coupling is common! refinement & application novelty & cutting-edge Regardless of your specific area of interest, unmanned systems researchers are challenged with operationally developing all five of these domains to a minimum level of operation. An unmanned system developed to this level can be called a core competency system.

6 Capabilities: WAM-V WAM-V Surface Vehicle 16 commercial OTS platform with enhancements. 450 lb holonomic propulsion system (electric) for full 3 degree-of-freedom motion. Mission-adaptable modular mounting rack for various instrumentation setups. Standard instrumentation: GPS, IMU, camera/vision, 900MHz/2.4GHz/5GHZ wireless. Robot Operating System (ROS) publish and subscibe software architecture. APIs in C++, Javascript, Matlab, Python. Unfused positional precision of +/-2.5m; station keeping and waypoint navigation to +/- 3m. Lidar-based SLAM (simultaneous location and mapping) for obstacle detection, avoidance and localization estimation.

7 Capabilities: other vehicles Unmanned Port Security Vessel (UPSV) 8 surface vehicle WAM-V + ROV launch & recovery system Mr. UH 4 surface vehicle

8 Capabilities: algorithms 2D lidar simultaneous localization and mapping (SLAM) Color thresholding

9 Capabilities: algorithms 3D scene reconstruction from vision Image classification using machine learning (regional convolutional neural network)

10 Past Work AUVSI Maritime RobotX Challenge 2016 Ordnance Reef WAM-V+ROV Integration 2017 Photos courtesy of RoboNation [1] (flickr) ARL Malia 2017 ARL Mantas 2018

11 Past Work images redacted for attribution, originals referenced here: AUVSI Maritime RobotX Challenge 2016 Ordnance Reef WAM-V+ROV Integration 2017 ARL Malia 2017 ARL Mantas 2018

12 Short term goals: RobotX 2018 The AUVSI Maritime RobotX Challenge is dedicated to advancing autonomous maritime systems (surface and underwater). Required autonomous functionality includes: Obstacle detection and avoidance Navigation and path planning Sonar-based navigation Docking Environmental manipulation Object (shape) detection and color recognition Sonar-based navigation Photos courtesy of AUVSI [2] and RoboNation [1]

13 Short term goals: RobotX 2018 The AUVSI Maritime RobotX Challenge is dedicated to advancing autonomous maritime systems (surface and underwater). Required autonomous functionality includes: Obstacle detection and avoidance Navigation and path planning Sonar-based navigation Docking Environmental manipulation Object (shape) detection and color recognition Sonar-based navigation images redacted for attribution, originals referenced here:

14 Long term goals Sensor fusion for localization Operation in GPS-denied environments: improving dead reckoning performance, and fusion of ground truth sensors. Non-conventional sensor fusion: feature tracking (vision), doppler velocity log, multi-beam sonar. Multi-robot, multi-domain, cooperative localization. Energy-efficient surface vehicles Renewable energy integration. Over-actuated, energy-optimized propulsion. Alternative propulsion methods. Cost-concious robotics Omni-scopic visual scanning Low-cost sensor fusion Solid-state lidar

15 For more information visit our website at: rip.eng.hawaii.edu