Autonomy and Behaviors
Beyond providing software that is flexible, operationally robust, and easy to use with user-designed payloads, Bluefin AUVs also demonstrate advanced autonomy and behaviors. Our approach to autonomy involves providing individual AUVs with three core capabilities:
- Flexible Mission Execution: The ability to execute mission plans in a flexible manner.
- Reliable Communications: The ability to robustly communicate information for effective situational awareness and command control.
- Enhanced Autonomy: The ability to build upon and expand core autonomous capabilities via a standard interface.
We believe that execution of mission plans should be sufficiently flexible to allow an AUV to tailor its actions in response to dynamic situations. This kind of autonomy reduces the burden placed on the operator by allowing vehicles to manage many of the details of mission plan execution and only occasionally request or require assistance. Bluefin’s next generation behavior control system provides a highly flexible executive responsible for accomplishing the goals specified in a mission plan. This behavior control facilitates dynamic insertion, removal, and modification of mission elements during execution. This is a vital capability in the often uncertain and noisy environments that AUVs face.
Behavior control is structured around behavioral capabilities, rather than low-level actions. For example, during mission planning the user simply specifies the region to be surveyed and the type of information required using our intuitive graphical Planner application. Mission plans are viewed as guidelines for execution. The mission is concisely represented in a plan file and sent to the AUV, which autonomously elaborates on the plan to have the right level of detail for execution and to be consistent with its capabilities. The same mission sent to a different vehicle might be executed slightly differently depending that that vehicle’s capabilities. This is all done without the operator requiring any detailed understanding of the inner workings of the AUV.
Bluefin’s vehicle autonomy is designed to work robustly in the face of limited, intermittent or unreliable communications. This robustness is provided by two core mechanisms: our communications infrastructure and the mission plan representation used by our behavior control. Our communications infrastructure provides different levels of reliability for messages, depending on their time criticality and frequency of transmission, and also appropriately filters those messages to fit the available bandwidth. At the same time, mission plans are designed to be compact and thereby easily communicated across bandwidth-limited, degraded, or noisy channels.
Our Standard Payload Interface provides a powerful mechanism for smart payloads to build on the core autonomy capabilities of our AUVs. The Standard Payload Interface works in concert with all of the vehicle software systems as well as with the Operator Tool Suite. Smart payloads can utilize raw or processed data streams from the core AUV to make decisions about what to do next, and communicate these decisions as requested modifications to the executing mission. The AUV processes and executes these changes using the behavior control system and other core software. Consequently, payloads cannot jeopardize the safety or integrity of the AUV. For this reason, smart payloads are sometimes referred to “backseat drivers” while the core vehicles systems are the “front seat driver”.
While some customers use the Standard Payload Interface for their own payload development, others use it to interface Bluefin AUVs with third-party autonomy capabilities such as MOOS-IvP. Researchers have developed a MOOS module that interfaces with Bluefin’s Standard Payload Interface to provide core vehicle data to the MOOS system while also allowing MOOS to operate as a backseat driver by issuing heading, depth and speed commands to the core vehicle. The use of MOOS-IvP opens the work of a whole community of autonomy researchers to Bluefin customers who wish to use MOOS as the basis for their smart payloads. For more information, visit the Sensor Integration section.
Download the Standard Payload Interface Specification to learn more.