My company, Overwatch, provides insruance for Unmanned Systems in the USA and Canada and globally. The challenge providing good risk financing (insurance) to any operation is to understand the risks associated and then to reduce these risks into measurable metrics which can be used to mathematically model the expected risk and then charge a proper amount of premium for the risk. All insurance works based on the law of large numbers so consequently it is easier to reduce the costs of insurance, such as auto insurance, when you have millions of users and buyers.
The current lack of FAA response to regulation, particularly as it applies to commercial drone use, has done little to stem the growth of the industry or the commercial use of drones. Just last week I was in a realtors office here in Texas and someone had just dropped off a flyer for aerial photography of homes, to be used on the realtor.com website for increase home sales, shot by a private UAV. Clearly this service was being sold for profit and clearly it is not authorized under the current FAA rules.
There are however many drone operations that may not be for profit that happen everyday that need insurance. Law Enforcement, Search and Rescure, First Responder, Research, Education are just a few. These operations can be properly supported as authorized under the current FAA rules.
One of the biggest problems providing insurance for drone operators is qualification. Normal commercial or private pilots have a long standing process of education, licensing and even physical qualifications in order to be a pilot. The advent of drones certainly changes these rules but since no standards exist now it can be very hard to determine what the right qualifications are to fly a drone and insure a drone operator on a system. Drone training courses are hard to come by. The U.S. Military appears to be one of the best non pilot training programs but most of us don't want to join the military just to get sufficient training.
You cannot seperate the issue of FAA regulations from the underlying risk issues of pilot qualifications, manufactuering qualifications and ultimately risk financing (insurance). It is impossible to put together "the rules" without some baseline minimum requirements for the participants.
Overwatch is putting a mark on the wall with our new UAV/UAS aviation liability product which we call the Remotely Piloted Aviation System (RPAS) policy. We are asking for this group's input regarding the coverage required, minimum Ground Station Operator pilot experience and even ideas on minimum safety standards for manufactured products. We need to have this broader discussion and we need to approach the FAA with proposed solutions vs. waiting for solutions to be provided during a time of government furloughs. The private sector and individual users are the best group to submit intelligent plans and ideas. I say this with confidence because it was private sector logistics companies that came up with the idea of the Northern Distribution Network (NDN) that allowed the logistics resupply of the US Military in Afghanistan via a Northern Route avoiding Pakistan. Many said it couldn't be done. Many in the military said it couldn't be done but the private sector believed it could be done and they did it allowing the sustainment of ground forces in Afghanistan.
If you would like to post your input here we will do our best to incorporate your input into future products and policies that support the entire community. Overwatch will be at the AUVSI show in August in Washington, D.C. and also at the Canadian Unmanned Systems show in Vancouver in November. Please stop by and share your thoughts with us if you can.
Replies
Grant,
This issue is very relevant for serious companies who want to build a more interesting commercial future using the technology. The tech is only a means to an end though, a means to collect data usually which is later post-processed to deliver the actual value.
I've just added a discussion on the issue of templates and records that are probably going to be needed for purposes like cost control, but I reckon also for assessments in an insurance claim and after a larger period could become a basis for determining sharper premiums.
In Europe there's the EU unmanned systems centre, largely a UK effort, which looks at issues like insurance and legally making use of the airspace and details the legal requirements. They further subdivide the aircraft by weight, which is pretty useful: http://www.eurousc.com/
CASA has developed some forms for "Operator's Certificates" for drones specifically: http://www.casa.gov.au/scripts/nc.dll?WCMS:STANDARD::pc=PC_100377Se... for "UAV operator license" on google shows a couple, but not all of them are actually accredited.
Canada has the procedure for SFOC here: http://www.tc.gc.ca/eng/civilaviation/regserv/cars/part6-standards-...
Some more academic info to "manning" uav's: http://www.airpower.maxwell.af.mil/airchronicles/apj/apj05/spr05/ho...
Flight school: http://www.resource-uas.co.uk/Home.aspx
As for your questions:
As this area is pretty new and the current list of "best practices" is not very extensive, risk management should be an integral part of the organization undertaking such activities, beyond policies for safety management (procedures in place to prevent harm to operator, third party bystanders, equipment, etc... etc... ).
My most immediate concern is damage to third parties. In my case the airframe and electronics aren't that expensive to replace, they're certainly not the same value of a car, so insuring against such a loss is almost silly. One or two flights would make up for it (this isn't true for everybody though, since some aerial filming guys use very expensive cameras onboard and their ROI times are more like 60 flights).
I reckon that third party damages are going to make the biggest numbers here and this very highly depends on the area flown. For example, in the middle of the Amazon it doesn't make sense to cover for third party damages, but ecological damages can / may exist (which is hard to quantify). It is a tricky one, because the monetary consequences are usually very small, except when the crash results in a lipo fire setting a very dry region of forest alight. The operator won't have the resources available to extinguish that and if that fire spreads over a wide region, the damages are enormous. ( thus, this feeds back into the risk management and flight preparation stages ).
There are situations in which flights over small town villages ( 500/1000mx500/1000m ) are desirable.
One of the issues in UAV developments is that the equipment basically flies itself. The human is kept necessarily in the loop to have a focal point of final responsibility (just like in regular aviation). In regular aviation however there are serious concerns with life onboard the aircraft as well. From an electronic and software perspective, it becomes very important to have analytic software in place that can detect failures. Most autopilots rely on sensory information without cross-checking with other systems and this can have disastrous consequences. It is therefore up to the pilot to confirm the plane isn't doing anything "funny". There are measures like parachutes that can significantly decrease consequential harm/damage if deployed automatically in the case of a failure.
Paparazzi and arducopter both have measures in place which cordon off a specific area that the uav is supposed to stay within. In the first boundary outside this area the uav is told to fly home. If it reaches the second boundary it is told to take a dive and power off. These measures help to prevent fly-aways that go over a large distance, because some aircraft can now stay in the air for 50+ minutes, which covers miles.
There's also a distinction between RPA and autopiloted aircraft. If flying through a video and control link directly then any external influence causing the link to be broken (object moving in RF path, radio/radar turning on, a wifi turning on close by) has immediate consequences. For the autopilot they're not immediate, but autopilots need to deal with an immense scala of flight dynamics and the failure of a single sensor can have disastrous consequences, where an RPA pilot would simply rely on other information like the visuals to safely land. The difference in risk and probability of failure between the two needs to be analyzed.
I'm also with you regarding components. There are already existing certifications with regards to emission ratings. These emissions could hamper the ability of a receiver to receive the control signal. Since a component with a high amount of emission is within 20cm vs. the control radio at a distance of 500m or larger it's easy to understand that at some point such a noisy component is going to overpower this control link.
There's also the issue of flight preparation prior to flight and pre-flight checks. This ties in with safety management principles like checklists, verifying correct functioning and so on.
A number of accidents have occurred caused by transport, where a bit of rough road dislodged or impacted some equipment on board which later caused an equipment failure.
There's also the issue of making sure the operator is safe during the operation. Some people reported flying with goggles on in FPV whilst standing in an ant hill or close to a beehive. So it's not just about the safety of the vehicle, but also the safety and comfort of the operator during the operation that needs to be guaranteed.