Previous Page  18 / 34 Next Page
Information
Show Menu
Previous Page 18 / 34 Next Page
Page Background

ACT Autonomous Surface Vehicle Workshop Report

Page

18

of

34

In this scenario, with no warning systems and only manual operation of sensors, an

organization would likely operate with an “attended” level of supervision. Operational

sensing systems such as video, radar or automatic identification system would require

continuous manual monitoring and payload sensors such as sonars and navigation

systems would require careful scrutiny to ensure successful data collection and to start

and stop logging of sensor data when desired.

Now assume the same scenario, in which the vessel is enhanced with a Level 2 Self-

awareness, being able to log parameters about its own systems and to provide operator

warnings when systems go awry. Consider also that the sensor autonomy level is

enhanced to Level 2, now being able to start and stop sensors automatically at each

point within the mission. The vehicle is now able to warn the operator about faulty

systems, low battery voltages, over currents and other warnings, freeing up one’s

attention to focus on other things. Moreover, because these systems record a history of

normal operation, an operator can scrutinize this history for common faults and

conditions leading to malfunction. That history can then be used to anticipate faults and

failures allowing one to mitigate their risk to mission success.

In addition, one need not manually manage starting and stopping of logging of sonar

and other sensor data. At these levels of autonomy, the operation begins to feel

something like a standard hydrographic survey launch. In this model of operation, the

autonomous coxswain is “under-instruction”, that is, reasonably autonomous but

unable to make decisions on his/her own. Level 2 Sensor autonomy puts in place a

system not unlike the commercial data acquisition software Hypack; able to turn sonar

data logging on and off at the beginning and end of each survey line respectively

without operator action. A survey technician must still scrutinize the data and a

knowledgeable operator must still provide situational awareness and directive to the

autonomous coxswain when warranted.

As one gains experience operating in an area, learns common traffic patterns,

understands a particular vessel’s likely faults and character, or, if one has a higher risk

tolerance in general, one might operate in a monitored level of supervision with these

levels of autonomy. The vessel would largely drive itself from waypoint to waypoint with

few obstacles or other vessels with which to contend. If sensors can be statically

configured and still collect quality data without readjustment, such operations might be

successful. However, with no self-aware model of operation and no ability to monitor

systems for expected behavior (Self-awareness at Levels 3 and 4), operators who will be

forced to hope that nothing unexpected occurs. The vehicle will be unable to warn when

propulsion systems are fouled, when the vessel snags on fishing gear or when navigation

systems begin to provide estimated positions rather than measured ones. Therefore,

while monitored operation is possible, is requires a considerably higher risk tolerance.

A common mode of operation under consideration is colloquially known as the

“mother-duck and duckling” operation in which a manned surface vessel conducting

hydrographic survey operations is flanked by one or more autonomous vessels that

operate a fixed range and bearing from the manned ship with their own hydrographic