UAS – Study ‘Sense and Avoid’ as a Replacement of ‘See and Avoid’

  • Home 2009 UAS – Study ‘Sense and....

UAS – Study ‘Sense and Avoid’ as a Replacement of ‘See and Avoid’

48TH ANNUAL CONFERENCE, Dubrovnik, Croatia, 20-24 April 2009

WP No. 91

UAS – Study ‘Sense and Avoid’ as a Replacement of ‘See and Avoid’

Presented by TOC

Summary

Before unmanned aircraft can be integrated into non-segregated airspace, they need to be able to comply with all applicable rules and regulations. This paper explains how ‘sense and avoid’ procedures could be used by Unmanned Aerial System (UAS) to meet ‘see and avoid’ requirements. IFATCA and IFALPA policies are presented, new terms from ICAO are introduced, and relevant rules and ‘sense and avoid’ technologies are examined. The paper is informational in nature; however draft recommendations are made to update Policy to reflect the new terminology.

Introduction

1.1  At the 2005 IFATCA conference in Melbourne, the IFATCA Technical and Operations Committee (TOC) presented a working paper entitled “Investigate operational use of Unmanned Aerial Vehicles“ (working paper 90). In that paper TOC concluded that:

“The problem of detecting, sensing and avoiding other aircraft during flight is a crucial challenge that must be overcome before civil aviation authorities permit UAVs to fly in unrestricted civil airspace”

and it was recommended that TOC be assigned a work item to study how UAV operators intend to deal with “see and avoid” requirements.

1.2  The ‘See and Avoid’ principle is a fundamental rule of flight which requires that when weather conditions permit, pilots must actively search for conflicting traffic with the object of avoiding collision. Developments indicate that Unmanned Aircraft Systems (UAS) ‘Detect and Avoid’ procedures could replace ‘See and Avoid’. For the 2008/09 work program, TOC was tasked with investigating whether ‘Detect and Avoid’ might allow the operation of UAS in non-segregated airspace; this paper is the result of that study.

Discussion

2.1 IFATCA

2.1.1 IFATCA Policy is:

“All Unmanned Aerial Vehicles operations in non-segregated airspace must be in full compliance with ICAO requirements.”

 


2.2 ICAO

2.2.1  ICAO has formed the Unmanned Aircraft Systems Study Group (UASSG), which is charged with coordinating the development of Standards and Recommended Practices (SARPS), Procedures and Guidance material for civil unmanned aircraft systems, to support a safe, secure and efficient integration of UAS into non-segregated airspace and aerodromes. It will be some time before the UASSG produces final recommendations to ICAO, however three terms that have been adopted by the UASSG and which will be used throughout this paper are detailed below.

ICAO UASSG draft recommendations, chapter 4 – ATM:

“TERMS, ACRONYMS, DEFINITIONS

Unmanned Aircraft System (UAS) – comprises of one unmanned aircraft (UA), control station (s) and any other elements required for flight.

Unmanned Aircraft (UA) An aircraft that is designed to operate with no human pilot on board.

Note. – The vehicle element of a UAS is an “aircraft” as defined in ICAO Annex 2, even though it is defined separately as “unmanned”.

See and Avoid / Sense and Avoid / Detect and Avoid – Since these terms do not currently appear in ICAO documents, it is assumed for the purposes of this document, that the ability to “see and avoid” relates to a human capability and “sense and avoid” or “detect and avoid” (used interchangeably) relates to a “technical” capability. Ultimately, it is still the pilot-in-command who is responsible for “detecting and avoiding” collisions.”

IFATCA has policy using the term Unmanned Aerial Vehicle. Policy using this term will need updating to use UAS or UA as appropriate.

2.2.2  ICAO Annex 2 Rules of the Air states:

“3.2 Avoidance of collisions

Nothing in these rules shall relieve the pilot-in-command of an aircraft from the responsibility of taking such action, including collision avoidance manoeuvres based on resolution advisories provided by ACAS equipment, as will best avert collision.

Note 1.— It is important that vigilance for the purpose of detecting potential collisions be exercised on board an aircraft, regardless of the type of flight or the class of airspace in which the aircraft is operating, and while operating on the movement area of an aerodrome.”

Other than referring to the use of ACAS, this rule does not explicitly state how the pilot- in-command should determine that there is a need to avert collision. A pilot could thus “see” or sensors could detect potential threats and in either case avoiding action is taken as required.

This rule could present a difficulty for UAS in that it is the pilot-in-command who has the responsibility for avoiding collision. In ICAO Annex 2 Rules of the Air, pilot-in-command is defined as:

“The pilot designated by the operator, or in the case of general aviation, the owner, as being in command and charged with the safe conduct of a flight.”

This definition does not by itself indicate that a pilot must be on board the aircraft; however the UAS may be operating in an autonomous mode without a pilot-in- command actively flying the aircraft at the time an avoiding manoeuvre is required. Indeed, for a number of reasons covered later in this paper, it is likely that detect and avoid technologies will rely upon the UAS automatically taking avoiding action.

2.2.3  ICAO Annex 2 Rules of the Air also states:

“3.2.2.3 Converging. When two aircraft are converging at approximately the same level, the aircraft that has the other on its right shall give way, except as follows:

a) power-driven heavier-than-air aircraft shall give way to airships, gliders and balloons;

b) airships shall give way to gliders and balloons;

c) gliders shall give way to balloons;

d) power-driven aircraft shall give way to aircraft which are seen to be towing other aircraft or objects.”

In order to determine that it is necessary to give way, the power-driven aircraft needs to sight any converging airship, glider, balloon or aircraft that is towing. This implies a UAS needs an optical sensor and that it has the capability to process the data provided by this sensor, so that the appropriate action can be taken.

2.2.4  Finally, ICAO Annex 2 Rules of the Air states:

“3.2.5 Operation on and in the vicinity of an aerodrome

An aircraft operated on or in the vicinity of an aerodrome shall, whether or not within an aerodrome traffic zone:

a) observe other aerodrome traffic for the purpose of avoiding collision;”

In using the word ‘observe’, the implication once again is that a UAS which is to operate at a non-private aerodrome requires an optical sensor in order to comply with this rule.


2.3 IFALPA policy

2.3.1

“IFALPA ANNEX 8 APPENDIX ‘AIR-C’

Preamble

The IFALPA policy on Unmanned Aerial Systems is intended to protect and enhance aviation safety to the highest standards by promoting a single level of safety worldwide for all users of civilian airspace.

IFALPA believes that UAS technology is not capable of replacing human capabilities, particularly in complex and safety-critical situations. Therefore, IFALPA strongly opposes the use of UAS to supplant the role of pilots in any type of air transport operations.

IFALPA believes that the safe integration of UAS operations into civil, nonsegregated airspace can only be achieved if these aerial Systems are regarded in all ways as an aircraft, and they and their operations must comply with all existing rules and regulations applicable to the same class of manned aircraft.

Non-compliant UAS will require segregated airspace.

IFALPA believes that it is not acceptable for existing rules and regulations to be changed in order to integrate these aerial systems and their operation.

Note.- The special characteristics of theses systems and their operations have to be taken into account i.e. Oxygen systems, windows.

To achieve the above, IFALPA believes that the following items have to be addressed.

Note.- The list is not limiting.”

The list of items referred to above is draft policy and from this list:

“2. Air Traffic Control considerations

2.1 As far as Air Traffic Control and other air transport participants are concerned, a UAS should behave like a manned aircraft and be subject to the Rules of the Air. The operation of UAS in civilian airspace should not make any difference (for example through special flight procedures) to the daily operation of other air traffic participants (commercial and general aviation).

2.2 Each UAS must have a designated Commander (Pilot In Command) for the duration of the flight Operation, who shall ensure that the UAS complies with the Rules of the
Air.

2.3 The time delay between ATC instructions and vehicle response due to datalink/ communication transmission time delay should not exceed that for manned aircraft.

2.4 The problem of `See and Avoid ́ / `Sense and Avoid ́ has to be resolved.

2.5 UAS have to fit into the existing and future ATM environment and the generally accepted performance criteria for the environment they are operating in.

2.6 State-operated UAS should not be exempt from the above requirements.”

2.3.2 The IFALPA draft policy on Air Traffic Control considerations is compatible with IFATCA Policy. It is important to note that:

“IFALPA believes that it is not acceptable for existing rules and regulations to be changed in order to integrate these aerial systems and their operation”.


2.3 State rules

2.4.1  Some States do explicitly require a pilot to see and avoid conflicting aircraft. Such requirements are similar to this example from the United States of America (USA) Federal Aviation Administration (FAA) Rule part 91:

2.4.2

Ҥ 91.113 Right-of-way rules: Except water operations.

(a) Inapplicability. This section does not apply to the operation of an aircraft on water.

(b) General. When weather conditions permit, regardless of whether an operation is conducted under instrument flight rules or visual flight rules, vigilance shall be maintained by each person operating an aircraft so as to see and avoid other aircraft. When a rule of this section gives another aircraft the right-of-way, the pilot shall give way to that aircraft and may not pass over, under, or ahead of it unless well clear.”

2.4.3  The United Kingdom (UK) Civil Aviation Authority (CAA) has a policy on UAS sense and avoid, published in CAP 722 Unmanned Aircraft System Operations in UK Airspace – Guidance, Section 2, Chapter 2:

“5.1.1The overriding principle when assessing if a proposed UAS Sense and Avoid function is acceptable is that it should not introduce a greater hazard than currently exists. Any proposed function must demonstrate at least equivalence with manned aircraft safety standards and, where these standards exist, the UAS must comply with the rules and obligations that apply to manned aircraft including those applicable to separation and collision avoidance.

5.1.2 The separation and collision avoidance capabilities must be able to:

  • Detect and avoid traffic (air and ground operations) iaw Rules of the Air;
  • Detect and avoid all airborne objects, including gliders, hang-gliders, paragliders, microlights, balloons, parachutists etc;
  • Avoid hazardous weather;
  • Detect and avoid terrain and other obstacles;
  • Perform equivalent functions, such as maintaining separation, spacing and sequencing that would be done visually in a manned aircraft.”

2.5 See and avoid

2.5.1  See and avoid sits in the collision avoidance layer of conflict management. It is the one tool that is available regardless of aircraft equipment or air traffic service. Other tools, such as Traffic Collision Avoidance System (TCAS) and traffic information from a third party are often not available to the pilot, particularly when operating VFR or outside of controlled airspace.

2.5.2  The Australian Transport Safety Bureau (ATSB) research report (1991) into the “Limitations of the See-and-Avoid Principle” and other studies highlight that the ‘see and avoid’ principle is far from reliable.


2.6 Detect and avoid

2.6.1  A ‘detect and avoid’ system consists of optical, radar and other sensors which pass data to a processor that determines the likelihood of a collision and directs a course of action to avoid the collision. The processor may be an on-board system, or may be a human who is operating the UA remotely.

2.6.2  One major issue facing the UAS community is that of maintaining a reliable and fast communications link between the vehicle and its operator. There are problems with lack of sufficient bandwidth to carry the data, potential for interference and some degree of latency. With there being no guarantee that information on a detected collision threats can be relayed back to a pilot on the ground, a practical detect and avoid system will probably require a certain amount of autonomy, whereby avoiding action is taken without the input of the pilot.

2.6.3  The UK CAA suggests that a target level of safety for a ‘detect and avoid’ system is that it is at least equivalent with manned aircraft safety standards. However as the reliability of the ‘see and avoid’ principle is questionable, equivalence with manned aircraft may be setting the standard be too low. If the number of UAS increases rapidly, as has been suggested as likely for the smaller sizes for a variety of civil applications, the potential for collisions could be much greater than with manned aircraft today.

2.6.4  The technology behind detect and avoid systems is presently being developed and tested by many UAS manufacturers, while at the same time the Radio Technical Commission for Aeronautics (RTCA) Special Committee (SC) 203 is producing minimum performance standards for such a system. Manufacturers may gain FAA and other certification more easily if their products meet the RTCA standards.

2.6.5  Prior to its last meeting in October 2008, RTCA SC-203 had indicated that standards would not be available until 2019. It now says that it will produce standards by the end of 2013. It will achieve this by restricting these to the operation of UAS in class G and other airspace in which an ATC clearance may not be required, such as VFR operation in class E or F airspace. Standards for remaining airspace are still not expected until 2019; however contingencies for cases where a UAS enters controlled airspace will be included in the initial standard. The reduction in scope by RTCA SC-203 is designed to simplify the task in producing the standard and to produce a standard for the airspace in which most UAS operations are presently desired.

2.6.6  TOC finds it surprising that these standards are being developed for the very classes of airspace in which ‘detect and avoid’ is most likely to be used. For example, whilst ‘see and avoid’ might be required in class A airspace, it is very rarely used there and so class A airspace may have been a better place to start using a ‘detect and avoid’ system.

2.6.7  An example detect and avoid system is being tested by Northrop Grumman. This system is comprised of TCAS, an Automatic Dependent Surveillance – Broadcast (ADS-B) In sensor, a purpose built radar and an electro-optical sensor. The system is designed to handle up to 30 intruder aircraft simultaneously.

2.6.8  It is necessary to use a combination of sensors to aid detection of hazards. TCAS and ADS-B In are useful when the object is co-operative, that is another aircraft equipped with an operational transponder. Many hazards that a UAS could encounter, are however non-cooperative; these include birds, parachutists and even some types of aircraft.

2.6.9  On-board radar can be a useful sensor, but this does have its limitations. Many of the non-cooperative hazards have a small radar footprint. Also there is a minimum size below which a radar sensor is impractical. A possible solution is a Light Detection and Ranging (LIDAR) sensor; these have now been miniaturised and adapted for airborne use.

2.6.10  A large amount of work is being undertaken on the development of optical and infra-red sensors. These are required so that the UAS can detect hot air balloons, birds and other objects that are potentially invisible to radar. Under some circumstances a UAS may also need to visually identify an object, such as a glider that may have been detected by its radar, in order to determine which of the two have right of way. When flying under visual flight rules, a UAS may need to use its optical sensors to help it remain in Visual Meteorological Conditions (VMC).

2.6.11  An optical sensor is also required if a UAS is to have the ability to comply with many aerodrome operation requirements. Examples include not crossing illuminated stop bars and carrying out an instruction to follow or pass behind other traffic.

2.6.12  While the requirement for optical sensors exists, the reality is that the technology is not easily able to perform many of the functions noted in the two preceding paragraphs. The major difficulty lies in the volume of data that needs to be processed in real time in order to detect and identify an object.

2.6.13  An interesting area of research is that being carried out into acoustic sensors which could be used to listen for intruder aircraft. TOC is not aware of any examples of acoustic sensors which have been successfully tested in a ‘detect and avoid’ system.

2.6.14  Once a UAS has detected an object, the UAS may have to take action to avoid the object. As some UA, even those at high altitude, have relatively low energy, the manoeuvring time required could be quite long, with the UA potentially needing minutes to avoid a fast object such as a jet in the cruise.

Conclusions

3.1  Various new terms have been adopted by the ICAO UASSG:

  • The term Unmanned Aerial Vehicle is now simply Unmanned Aircraft (UA).
  • Unmanned Aircraft System (UAS) comprises the UA and any other system elements, such as a Ground Control Station and communications link necessary to operate a UA.
  • The term ‘sense and avoid’ has been replaced by ‘detect and avoid’.

IFATCA policy using the old terms should be amended to use the new terms.

3.2  See and avoid is neither defined, nor required by ICAO. ICAO rules do require that an aircraft takes all steps possible to avoid collision; these steps could include ‘detect and avoid’.

3.3  States which explicitly require a pilot to follow the ‘see and avoid’ principle may need to change their rules to include the ‘detect and avoid’ principle. Such a rule change may allow UAS with ‘detect and avoid’ capability to operate in non-segregated airspace.

3.4  ICAO rules make the pilot-in-command of an aircraft responsible for avoiding collisions. It may be necessary for the definition of a pilot-in-command to be changed to accommodate UAS operations.

3.5  Any ‘detect and avoid’ system for a UAS must provide an equivalent and ideally a much greater level of safety than the ‘see and avoid’ capability of human pilots.

3.6  Critical to the success of UAS ‘detect and avoid’ is optical sensing and processing. Optical sensor technology may have to improve considerably before a practical ‘detect and avoid’ system is developed. An optical sensor component of a detect and avoid system may allow the UAS to comply with the ICAO requirement for an aircraft to ”observe other aerodrome traffic for the purpose of avoiding collision” .

3.7  A ‘detect and avoid’ system cannot rely upon an operator’s intervention and thus may need to be capable of operating autonomously.

Recommendations

It is recommended that;

4.1  All references to ‘Unmanned Aerial Vehicle’ or ‘UAV’ in the IFATCA Technical and Professional Manual are changed to read respectively ‘Unmanned Aircraft’ or ‘UA’.

4.2  All references to ‘Unmanned Aerial Vehicles operations’ in the IFATCA Technical and Professional Manual are changed to read ‘Unmanned Aircraft Systems (UAS) operations’.

References

ICAO Annex 2 Rules of the Air.

IFALPA Annex 8.

FAA Rule Part 91.

CAP 722 Unmanned Aircraft System Operations in UK Airspace – Guidance (UK CAA).

Sense and Avoid Requirements For Unmanned Aerial Vehicle Systems Operating In Non-Segregated Airspace (NATO).

Limitations of the See-and-Avoid Principle (BASI, 1991).

Last Update: September 29, 2020  

April 14, 2020   905   Jean-Francois Lepage    2009    

Comments are closed.


  • Search Knowledgebase