The algorithms tested during 2012 LD-CAP flights were developed at the UNDUASE and MITRE. These algorithms use information from a multitude of sensors to determine if the current state and flight path of the UAS are conflicting with any other objects identified in the airspace. If a conflict is identified, then the algorithms send commands to the autopilot to avoid that conflict. This could be a change in speed or heading (horizontal maneuvers) or a climb/descend (vertical maneuvers) command.

Traffic Data

Generally, objects that are identified within the LD-CAP tests are cooperative and data regarding them are provided through an ADS-B data stream. In LD-CAP, only two aircraft participated in the tests (the NASA-Langley SR-22 and a UND aircraft). Other aircraft were avoided through full compliance with all Visual Flight Rules to ensure safety in the NAS.

Pilots in aircraft equipped with ADS-B In capabilities see aircraft with ADS-B Out capabilities within local proximity of the aircraft on an ADS-B display. In addition, the ADS-B data stream provides information regarding targets that are “pinged” by air traffic control secondary radars (MODE-C/S). This information is received by the ADS-B unit as TIS-B information.

TIS-B

One idea, the exploration of which is being led by MITRE, that is being tested is the conversion of data from primary radar returns (i.e., data regarding non-transponding aircraft) from air traffic control radars into primary TIS-B (PTIS-B) messages. There are challenges associated with this process. Secondary radar returns are essentially transponder returns from cooperative aircraft (the transponder on the aircraft is providing information about aircraft position to the radar when the radar scans the aircraft).

Within these data is aircraft barometric altitude (a key parameter in the ADS-B data stream). Because ATC radars are generally fan beamed radars (very poor vertical resolution but very fine horizontal resolution), primary radar returns provide relatively poor altitude discrimination (unless primary returns are correlated to ADS-B data or secondary radar returns—i.e., if an aircraft is already transponding). Procedures that will allow transmission of these primary returns in a similar fashion to that of secondary radar returns are being explored. Questions that are being explored include:

• Can one accurately create TIS-B messages from primary radar returns?
• How is primary radar information utilized by algorithms and pilots if available in the ADS-B data stream?
• Does lack of altitude information in PTIS-B messages cause problems with pilot situational awareness?
• How would geometric altitude (obtained with a pencil-beamed radar) be converted to barometric altitude (needed by the aircraft)?

UND Algorithms

The University of North Dakota AirBorne Sense and Avoid Algorithm (UND ABSAA) has been developed at the Unmanned Aircraft Systems Engineering (UASE) lab at the University of North Dakota since 2008 through a series of Department of Defense funded projects including the GPAR-RMS project. The focus has been on an airborne payload for small UAS using ADS-B as the initial sensor for resolving local aircraft traffic positions and using closest point of approach calculations (CPA) together with interval programming to automatically avoid local airborne threats.

Through further iterations the system has incorporated behaviors that increase capability such as four-dimensional avoidance, right-of-way (RoW) rules compliance, and terrain avoidance, all of which make the system better able to operate safely in the National Airspace System (NAS). Information regarding this algorithm is provided here. During LD-CAP, the algorithm was integrated into the NASA’s SR-22 surrogate UAS for flight testing.

Mullins, M., Foerster, K., Kaabouch, N., and Semke, W., “A Multiple Objective and Behavior Solution for Unmanned Airborne Sense-and-Avoid Systems,”
AUVSI’s Unmanned Systems North America 2012, Las Vegas, NV, August 5-9, 2012.

MITRE Algorithms

MITRE has several variants of a sense and avoid algorithm that are being tested during LD-CAP. These algorithms were developed to enable testing the use of ADS-B data in ABSAA.

2013 Algorithms

The goals for 2013 are to test these algorithms in more dynamic environments (e.g. turn into conflict, both aircraft maneuvering, etc.). There are plans to include additional algorithms from other entities for the 2013 test flights, including Draper Laboratories.

Sim-to-Flight (MITRE)

Prior to flight testing, these algorithms were tested at MITRE in a simulation environment to understand how they would perform during specific encounters. These simulations also aid algorithm safety analysis. Once algorithms pass the simulation tests, they can be installed on computers within NASA-Langley’s SR-22. Scripted encounters are then tested in real flights to mimic those examined in simulation. This Sim-to-Flight capability was developed and streamlined during LD-CAP. More information on this capability is provided here.