Prev-Talk | Next-Talk | All-Talks | Talks-Sorted
Talk-02: Physics-based Simulation Environment for Adaptive and Collaborative Marine Sensing with MOOS-IvP
Henrik Schmidt, MIT, Laboratory for Autonomous Marine Sensing Systems, Center for Ocean Engineering
One of the main obstacles to the development of adaptive and collaborative autonomy is the unpredictability of the platform actions when the underwater vehicles are released to autonomous control based on real-time sensory input. The most efficient and cost-effective approach to mitigating the risk is the extensive use of a simulation environment which accurately mimics the actual system, not only in regard to the onboard autonomy and platform dynamics, but even more critical, accurately reproduces the performance of the sensor systems. For adaptive and collaborative acoustic sensing, a particular challenge in this regard is the real-time generation of acoustic sensor data that accurately represents the complex environmental acoustics of the ocean environment. With its modular structure, MOOS-IvP is ideal for establishing a multi-platform simulation infrastructure, where the actual autonomy configuration can be operated in a virtual environment which accurately mimics the real ocean. The Laboratory for Autonomous Marine Sensing Systems has developed a suite of sensor simulations of various levels of fidelity that allow the execution of highly realistic, adaptive and collaborative acoustic and non-acoustic sensing missions. The extensive use of virtual experiments in this simulation environment has been a key to a string of successful field deployments of distributed sensing systems over the last decade [Work supported by ONR, and the NATO Undersea Research Centre].
DOWNLOAD the brief given at MOOS-DAWG'11.
Categories:
- MOOS-IvP
- Ocean/Autonomy Simulation
- Ocean Modeling
- Autonomy / Collaborative Autonomy
- Mine Countermeasures
- Anti-Submarine Warfare