The introduction of new CASs results in a multiplication of the number of different systems in the airspace. And this is not without consequences. During ACAS Xa integration tests, though the coordination was working perfectly, it was noted that the performances of TCAS were degraded. This highlighted the fact that CASs need not only to be able to coordinate but they also need to be interoperable.
The automotive world is on the verge of a revolution with vehicles reaching higher levels of automation. Automation is defined by the Oxford English Dictionary as: “the use of electronic or mechanical devices to replace human labor”. Among all the existing concepts (e.g., Google, Uber, Tesla, but also Renault, Toyota and many more) the automation level can vary greatly. Indeed, different concepts have different goals, some of them aim at removing the human from the dynamical parts of the driving task, while others want to keep the human “in” or “on” the loop (e.g., cars with or without steering wheels). This created the need for a classification of the different goals and levels of automation, which led to the birth of the “J3016: Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles” document (can be bought obtained for free here). This document is the result of a common US/Europe effort to create a descriptive and informative but not normative document, which writing has been entrusted to the SAE on-road automated vehicle standards committee.
The main contribution of this document is the determination of six levels of automation, along with precise definitions for several terms.
This document divides the act of driving in three stages: strategic (e.g., choosing the route, timing), tactical (e.g., motion inside the traffic), and operational (e.g. reflex reactions). It is pointed out that the automation levels considered do not deal with the strategic part. Numerous definitions are provided, here are a few of the most useful ones with simple definitions of our own (see the actual document for precise definitions):
- Dynamic Driving Task (DDT): what most people would call driving, involves the tactical and operational parts mentioned earlier.
- Dynamic Driving Task Fallback (DDT fallback): the system that deals with the situation when things go wrong. Can be the human itself.
- Automated Driving System (ADS): the automation part of the car (hardware & software).
- Operational Design Domain (ODD): specific operational description in which an ADS will be designed to operate.
These four terms are the most used throughout the document, and in the main tables/figures. With these terms being defined, the document goes on to describe the six levels of automation that an ADS can provide. These are described in the following table.
– Level 0 = not automation at all.
– Level 1 = longitudinal or lateral distance handled autonomously.
– Level 2 = both distances handled, but human must continuously monitor the system.
– Level 3 = human does not need to monitor the system but he is the fallback system (must intervene if something wrong happens).
– Level 4 = human not needed anymore (even for fallback), but operations subjected to some limitations.
– Level 5 = like level 4 but with no operational limitations.
These six levels allow to describe precisely the levels of autonomy of existing cars and of the ones to come, for example Tesla cars are Level 2. Some more interesting concepts are detailed through the document, like the possibility to have a remote driver, or a dispatcher capable of initializing or deactivate the ADS. The monitoring (permanent vigilance) and receptive (no vigilance but can be alerted) concepts are detailed. The notion of minimal risk condition, some sort of contingency procedure, is also discussed. And more notions that would not fit in this short article.
To conclude, this document is not a standard, as it is repeated throughout the text; it is a descriptive and informative document. Yet, it is very detailed and clear, with numerous examples to illustrate the various concepts. From all the concepts, it appears that a parallel with aeronautics could be done, for example a “minimal risk condition” looks a lot like a “contingency procedure”. Defining a dictionary of equivalent notion to translate notion from one world to the other would appear highly useful. Especially as drone levels of automation are not yet clearly categorized and could use such a taxonomy, at least as a starting basis.
According to the ICAO’s RPAS manual, a full Detect And Avoid (DAA) system must prevent collisions with: conflicting traffic, terrain and obstacles, hazardous meteorological conditions, ground operations and other airborne hazards (such as wake turbulence, birds and volcanic ash). However, most of the existing efforts focus on DAA for conflicting traffic as it represent the highest risk, letting aside the rest of the hazards. Especially in the case of ACAS Xu which design and evaluations focus on conflicting traffic avoidance.
Recently, Trustwave applied for a patent describing how to integrate existing terrain and weather avoidance systems with ACAS Xu. The goal being to inhibit collision avoidance maneuvers which could direct the RPAS into terrain or hazardous weather, and to account for these in the computation of Remain Well Clear (RWC) maneuvers.
The efficiency of such a system remains to be demonstrated, yet it is one step closer to a complete DAA system.
The Airports Council International (ACI) recently published a position paper on Drone Technology giving an insight on their vision of the future. In this documents they acknowledge the important role that drones can play for the development of airport activities, the impact that drones traffic will have on airports, as well as the risks in termes of security and disruption of airport services.
The ACI asks for a common european effort, with a “no airport left behind” approach, and calls for cooperation with airlines, ANSPs and authorities, on topics including: the definition of restricted zones (geofencing), the detection and neutralisation of drones, and the definition of roles and responsibilities of the various actors. In this regard it strongly supports the U-Space initiative led by the SESAR-JU.
In terms of actions, the ACI World set up a “Drones Working Group” aimed at writing a Handbook and global guidelines for airports. At the same time, ACI Europe asks the EASA to write and publish a “European Safety Rulebook” to disseminate good practice and safety culture to the public. The ACI also acknowledge that a medium to long term integration will require to update relevant ICAO documents.
The envisioned roadmap for drones integration is to integrate the less risky operations as fast as possible, then define standard scenarios to enable operations in the EASA framework and finally gather from the aviation industry best practices and operational concepts.
In all the previous aspects, the ACI insists on the fact that any development must be “future proofed”, it is to say that it should be able to evolve as the technologies evolve.
The RTCA Drone Advisory Committee (DAC) is a committee aimed at supporting the FAA on their regulatory effort to enable drone integration in the national airspace. The 8th of November, the DAC is meeting to consolidate their finding and reach consensus on the recommendations to provide to the FAA. This is likely to trigger from the FAA an update of existing regulation thus impacting the whole drone industry.
More information here.
Studies for drone regulations accelerated the pace for the assessment of risk for drone operations. A recently published ‘Annual Safety Review 2017’ discusses the aviation accidents in detail containing a chapter specialized for drones. This report by EASA, involves the data from European Central Repository (ECR) experienced by EASA member states.
With the increase in the number of drones and possibly raising consciousness on reporting occurrences, the numbers of non-fatal accidents raised by 470% in 2016 relative to 2011-2015 average, luckily maintaining zero fatalities. Most of the times, it is commercial airliner pilots to report the occurrences, and rarely the UAS pilot.
The prior key risk areas has been investigated and aircraft upsets is by far the most common cause of the occurrences and set as the first key risk to address for safe integration of drones into airspace. 50% of RPAS accidents falls in this case which often results in a damage or destruction of UAS following loss of the control of the drone by the pilot.
Second key risk area is airborne collision although it is rarely encountered due to probable frequency with exponential increase in the number of drones. Obstacle collision is the 3rd risk area which will tend to increase with integration of drones especially in urban areas.
Ref : https://www.easa.europa.eu/system/files/dfu/209735_EASA_ASR_MAIN_REPORT_2017.pdf
The air traffic can be divided into cooperative and non-cooperative traffic. The cooperative traffic is equipped with avionics facilitating its detection. The non-cooperative traffic has no such equipment and detection is solely based on ground or onboard sensors. It is important to note that detecting cooperative traffic is a lot easier and more precise than detecting non-cooperative traffic. This is why many experts advocate for all low altitude traffic to be cooperative, at least in high density airspaces, and a proposed solution is to use ADS-B. This solution seem acceptable considering that a large part of the existing traffic is already required (or will be soon) to carry ADS-B out, the technical solutions exist and they are affordable both in terms of SWaP (Size, Weight and Power) and cost. Now, a crucial questions remains: is it possible to introduce hundreds of ADS-B users in already busy (radio frequency wise) airspaces without disturbing the performances of existing systems, e.g. ATM systems ?
To answer this question, the MITRE conducted a study on the impact of equipping low level drones with Universal Access Transceiver (UAT) ADS-B. Both air-to-air and air-to-ground communications were considered. According to this study the crucial parameters are the traffic density and ADS-B transmission power. The following table, extracted from the study, shows the probability to decode a message depending on drones density and transmission power with values in bold being acceptable for ATM applications. With a density of 5 drones per square kilometer the emission power cannot be higher than 0.01W, which strongly limits the communication range, though experiments to know the precise range depending on the transmit power should be conducted.
Overall, the results of this study show that using ADS-B UAT in high density airspaces will prove difficult has reducing the transmission power of ADS-B is likely to decrease detection ranges and impact safety. For the particular case of UAT, considering the fact that it is only used in the US, principally aimed at General Aviation (GA) and with the current grow in GA traffic, the FAA is unlikely to approve such solution to make the drones cooperative. From a broader perspective, the study showed how quickly a cooperative method can overload a communication mean. Having only cooperative traffic is desirable but this kind of study make it look like an unreachable objective. For now…