Detection, Characterization, and Evaluation of UAP: Opportunities and Challenges
Michael Lembeck, Ryan Graves, Hank Owen, Matt Miller, Ravi Starzl, Michael Greene
Unidentified Anomalous Phenomena (UAP) identification - a responsibility and opportunity for the technical community
Congressional committees, advocacy organizations, and scientific partners are continuing to raise awareness about Unidentified Anomalous Phenomena (UAP). The Department of Defense All-Domain Anomaly Resolution Office (AARO) specifically defines UAP as “objects that are not yet attributable to known actors and that demonstrate behaviors that are not readily understood by sensors or observers.” The Office of the Director of National Intelligence (ODNI), in their June 2021 UAP Preliminary Assessment characterized the ability of UAP “to remain stationary in winds aloft, move against the wind, maneuver abruptly, or move at considerable speed, without discernable means of propulsion.” Without an understanding of their origin or intent, these abilities potentially represent a threat to aviation safety and national security.
In light of these observations, the AIAA UAP Integration and Outreach Committee (“AIAA UAPIOC”) was formed to provide technical subject matter expertise spanning multiple fields relevant to the UAP issue. We have recruited experts in phenomenology, sensors, machine learning and other aerospace and scientific related disciplines to address this topic. This team has come together to discuss this challenge in the following thought leadership article.
It is possible to detect UAP, but the absence of broad-based collection efforts yields inconsistent data
The UAP encounters by the U.S Navy off the west and east coasts of the United States have demonstrated that it is possible to detect UAPs. In 2004, The USS Princeton with an AN/SPY-1B radar, and F/A-18 Super Hornets with upgraded AN/APG- 79 radars were able to detect and track unidentified objects in the airspace of their military training areas. In addition to these military reports, commercial airline pilots have been reporting multiple recent physical sightings of UAP over the Pacific Ocean and continental United States. Unclassified UAP related sensor information can be found and are updated on the All-Domain Anomaly Resolution Office’s website under “UAP Videos.”
The utility of available UAP data is challenged by the lack of an open and consistent reporting and data collection process, resulting in data sets that are too limited to allow for detailed trend and pattern analysis. Sparse sensor vantage points rely on the coincidence of the phenomena’s appearance to make observations. UAP data, when released into the public domain without a vetted and well-understood source accompanied by appropriate metadata, often raises chain-of-custody concerns that impacts the assessed legitimacy of the event.
A sequential, physics-based approach to UAP, combined with a structured data plan, can yield high quality data and insights
The enigmatic nature of UAP can benefit from a “bottoms up,” physics-based approach, where sensor phenomenology, sensor system calibration, and analytical models are specifically matched to known or suspected UAP characteristics in an iterative process. Once calibrated through a structured data collection plan, other technical and procedural issues can be mitigated. Finally, any findings of unique appearance, performance, and signature can be noted, and recommendations formulated to improve safety-of-flight in the proximity of such threats.
Accordingly, we propose an approach to further our understanding of objects of unknown origin operating in controlled airspace starting with a sequence of detection, characterization, and evaluation.
Detection – as part of a structured data collection plan, available sensing modalities are used to collect information on a variety of UAP phenomenologies, including their physical configuration and interactions with the surrounding environment.
Characterization – novel models are developed to cull UAP events from passive and active data tracks for further review. Algorithms and advanced signal processing techniques are applied to identify the capabilities and performance of detected UAP.
Evaluation – provide action plans and warning products in human usable formats for enhancing aviator safety.
Detection
An approach to Phenomenology, Sensor Systems, and Data Collection Plans
The utility of available UAP data is challenged by the lack of an open and consistent data collection and reporting process. Ad hoc UAP reports are inconsistent in detail, limiting the detailed analysis of trends and patterns. To address this shortcoming, we are surveying commercial and public sensor networks whose mission objectives could be extended for UAP detection.
A key first step in the assessment of sensors that could be used to detect UAP is to enumerate the phenomenologies that anomalous objects might produce in their operations. Each phenomenology represents some physical interaction with the environment. Other limitations can impact the probability of detection, such as target range, atmospheric attenuation, or ionization shielding. As our understanding of UAP characteristics expands, physics-based requirements will be updated to match sensor compatibility with UAP phenomenologies.
A survey of known phenomenologies, sensors to detect them, and limitations in their application are provided below.
Relevant Sensor Phenomenologies for UAP Detection
Sensor Systems – limited by calibration and procedural constraints
Once relevant phenomenologies are identified, compatible sensor system can be selected from an inventory of purpose-driven systems that have been developed for tasks other than UAP detection. Some sensors are active in nature, transmitting a signal to detect non-cooperative or passive targets. Other sensors passively collect energy emitted by targets of interest. Some sensors are optimized to observe slowly varying phenomena, while others focus on quick measurements of targets with rapidly varying or high-performance behaviors.
By way of example, our group has collected an inventory of available radiofrequency (RF) systems that could be applied to the detection problem. These systems provide more accurate velocity and range information than other sensor modalities in many envisioned situations.
Relevant Radiofrequency Sensor Systems for UAP Detection
Data Collection Plans - enigmatic nature of UAP targets presents unique challenges for approaches using mature sensors
There are trade-offs that must be made in any data collection plan that employs mature sensors in an ‘off-brand’ manner. Optical military sensors are designed to detect vehicle heat sources such as products of combustion used in propulsion or aerodynamic heating of external surfaces caused by atmospheric friction. To repurpose these sensors for a function like UAP detection, some adjustments must be made.
How much work must be performed to adapt the output of the sensor to the UAP detection problem? The following figure qualitatively highlights examples in this trade-space.
Applicability of Using Existing Fielded Sensors in ‘Off Brand’ Manner for UAP Sensing
Commercial and open sensors such as the NEXRAD weather radar network has excellent coverage and data can be readily accessed. However, the readily available products are processed to be optimized for weather surveillance, removing point targets such as aircraft and other discontinuities. As shown in the NEXRAD Level I box, raw radar samples can be captured, and novel algorithms can be directly applied. Unfortunately, the slow scan rate, optimized for weather surveillance functions, is likely to add significant ambiguity when tracking highly dynamic targets. There is utility in using this sensor, but the threshold for doing so is relatively low, as the sensor is not optimized for this purpose. Other commercial radars may provide an opportunity to provide wide coverage at reasonable cost. However, modifications would have to be made to the processing so that initially a capture-then-process approach may be required.
Other passive radio frequency sensors, such as electronic warfare systems, could play a beneficial role in the passive detection of UAP. As in the case of NEXRAD, there is likely a tradeoff between the typical or default reports these systems provide in their intended role and a broader set of measurements that could be provided if raw data were captured for offline processing.
Tactical radars and air traffic control radars are likely to provide better track information for highly dynamic targets. These radars are widely deployed, and significant coverage can be expected from networks of these types of sensors. However, there are restrictions to accessing this data. Obsolescent tactical radars may be repurposed for other uses at a reasonable cost. When considering the total cost to acquire and field a radar system, the relative cost of elements such as the physical transmitter and antenna, especially if using an electronically steered phased array, typically are much higher than the signal processing functionality hosted on server hardware.
Two examples of the successful repurposing of tactical radar equipment include an AN/MPQ-64 Sentinel array used by the Naval Postgraduate School (NPS) Center for Interdisciplinary Remotely Piloted Studies (CIRPAS) to support meteorological radar research and an AN/SPY-1 transmitter and array used as part of the National Weather Radar Testbed in Norman, OK. Both were successfully used in unclassified environments by researchers to explore the application of this kind of relatively sophisticated radar technology to benefit meteorological radar advancement. A similar approach could be used to repurpose other tactical radars to set up relatively advanced UAP surveillance capabilities in an open, scientific environment at a relatively low cost.
Next Steps: Specific Data Collection Opportunities
Members of the AIAA UAP Integration and Outreach Committee have significant subject matter expertise, tools, and other resources to assist in the development of relevant UAP detection approaches. Data collection efforts that span a significant geographic region, are persistent in time, and offer multiple corroborating sensors is a challenging and sometimes costly effort. It is important to leverage existing systems and opportunities in a cost-effective manner. This approach can accelerate the time to fielding a useful sensor system.
The following figure highlights the trade-offs between observation persistence (the time that sensor can stay deployed) and the availability of multiple, corroborating sensors appropriate for the UAP challenge. Fielding many specialized sensors, for example, would be costly even for a relatively small geographic region and short time span. However, inappropriately leveraging large persistent networks may result in uncorroborated, ambiguous measurements.
Data Collection Event Persistence and Measurement Availability
A combination of military exercises, research opportunities, fixed and mobile systems, and commercial and military networks yields options in this trade space. Focusing on fixed locations or augmenting ubiquitous sensor systems with adjunct processing capability provides complementary approaches to UAP surveillance, addressing cases where long term, large area surveillance is required and cases where an expectation of UAP detection is higher allows for more corroborating and focused sensor employment. The table that follows highlights some examples for several collection opportunities.
Specific UAP Data Collection Approaches
Our Detection Experts
Hank Owen
HS Owen LLC / Lockheed Martin / U.S. Navy
Electrical Engineer, Radio Frequency (RF) Sensors and Propagation
Michael Greene
Ball Aerospace
Defense Systems Analysis
Characterization
Determination of UAP appearance, performance, and signature
AARO continues to review reported UAP events and has collated several trends that are appearing in the collected data. More than half of the reported UAP are spherical in shape. They mostly operate between 10,000-30,000 feet and can travel at speeds exceeding Mach 2. They are able to exhibit such performance without evidencing any obvious means of propulsion. The observation of such repeated patterns or signatures obtained from sensor data can be used to infer models characterizing UAP vehicle configuration and dynamics through several techniques and methods.
Our experts are familiar with the steps typically employed to develop such models and can apply the same techniques to the characterization of UAP.
Data Collection: Sensor signatures are first collected using various types of sensors. These could include radar, optical, or acoustic sensors. Each sensor type provides different types of data, such as shape, size, speed, acoustic signature, or heat signatures.
Signal Processing: The raw data from sensors is processed to filter out noise and extract relevant features. Signal processing techniques can help in identifying unique patterns or characteristics that are specific to certain vehicle types or their dynamics.
Pattern Recognition and Machine Learning: Advanced pattern recognition and machine learning algorithms can be employed to analyze the processed sensor data. These algorithms can identify specific vehicle types and configurations based on their unique signatures. For instance, the acoustic signature of a jet engine differs significantly from that of a propeller-driven aircraft.
Model Development: Once the type and configuration of the vehicle are identified, this information can be used to develop or refine models of the vehicle's dynamics. These models can predict how the vehicle will behave under various conditions based on its configuration. For example, the aerodynamic properties of an aircraft can be inferred from its shape and engine type.
Simulation and Analysis: The developed models can be used in simulations to analyze the vehicle's performance, behavior under different scenarios, and response to environmental factors. This step is crucial for understanding how the vehicle will behave in real-world situations.
Validation and Calibration: Finally, the models are validated and calibrated against real-world data or through controlled experiments. This ensures that the models accurately represent the vehicle's behavior. This process is widely used in military and civilian applications for tasks such as threat assessment, traffic management, autonomous vehicle development, and space vehicle design. It allows for a deeper understanding of how different vehicles will perform in various environments and conditions, which is critical for planning, design, and operational decision-making.
This process is widely used in military and civilian applications for tasks such as threat assessment, traffic management, autonomous vehicle development, and space vehicle design. It allows for a deeper understanding of how different vehicles will perform in various environments and conditions, which is critical for planning, design, and operational decision-making.
Novel models based on limited data sets
Characterizing any new target type is a challenging task, especially with the limited availability of useful data sets. Two general approaches are described here and form a basic framework from which to proceed.
First, in a model trained on signal characteristics, there is the opportunity to define parameters that can be associated with signal characteristics captured from UAPs. In general, the ability to discriminate one type of target from another benefits from the availability of a larger set of parameters, and also from the separability of target types as defined by a set of parameters being measured. For UAP these parameters may be novel, or they may have been used previously in characterizing conventional vehicle behavior. With this approach, we attempt to confirm a descriptive set of parameters and determine whether novel signal characteristics can be used to identify differences between UAP targets and conventional vehicle targets.
Second, in a model trained on background properties we may not be able to define new parameters or ranges of parameters that are associated with UAPs, but instead discriminate what is common or ordinary from anomalous behaviors. In this second approach, we define background properties and work to identify any novel tracks that deviate from this background. Both techniques can be useful, and they may be used in a coordinated or sequential manner as better characterization of UAPs is developed.
UAP Characterization Model Training Approach and Application
Novel Approaches to UAP radar signature simulation may be needed for effective characterization, given potential UAP signature management capabilities
The 2021 UAP Task Force Preliminary Assessment postulated that a small subset of UAP demonstrate a degree of “signature management,” that deliberately modified electromagnetic signatures to reduce the likelihood of detection. That attribute, combined with the inherently enigmatic nature of UAP appearance and performance, may require novel, iterative approaches to characterization.
High quality optical data can be used to estimate shape models of the imaged target. This information can be used as input to radar signature simulations iterating through surface shapes and material properties to get a match. For radar detections, modern computational electromagnetic (CEM) simulation tools can generate highly accurate radar signature data for targets if accurate shape and material data is available. For UAP, neither is openly available.
Given this lack of detailed information, one can consider an alternative approach that attempts to match synthetic radar signature data with measured data. A baseline target model for the UAP informed by imagery and other data sources can be generated and radar signature data produced for this initial model. CEM tools can rapidly produce fully polarimetric data for many aspect angles and frequencies. This data can then be compared against measured radar signature data such as raw radar cross section data (i.e., I/Q data), high range resolution (HRR), and synthetic aperture radar (SAR) imagery. The advantages of HRR and SAR imagery are that target scattering centers can be identified and correlated with target geometric features. An example of correlating geometric features with scattering centers in HRR data is shown in the image below.
This approach is iterative in nature and can be guided by additional AI/ML – driven non-cooperative target recognition techniques to iterate the target CAD model and material properties until good agreement is achieved between the synthetic and measured radar signature data. Such an approach can also account for uncertainties in the target orientation and target motion.
Another potential advantage to using simulated data is that one can begin to explore trade spaces for new radar systems. For example, we may be able to elucidate enigmatic aspects of UAP by increasing bandwidth, moving to fully polarimetric radar systems, and using non-traditional radar frequencies.
Our Characterization Experts
Matt Miller
Consulting
Radio Frequency (RF) Engineering
Clayton Spann
Spann Consulting Services (SCS)
Radar Cross Section (RCS) Engineering
Ravi Starzl
Carnegie Mellon University
Machine Vision / Machine Learning
Evaluation
Providing operators with actionable, safety enhancing information
After detecting and characterizing UAP operating in controlled airspace, the ultimate goal is to provide guidance to provide aviators with maneuvering advice and avoidance mechanisms enhancing the safety of flight. Ultimately, detection, characterization, and evaluation may require the practical integration of multiple sensors, networks, and enhanced processing via artificial intelligence/machine learning techniques to convert a detected UAP signature into aviation safety-enhancing warning products. Key elements of the evaluation process include:
Real-Time Data Processing: Once a UAP is detected, the data from various sensors must be processed in real-time on the platform. This involves filtering out false positives (such as birds, drones, or weather phenomena) and accurately characterizing the UAP in terms of speed, trajectory, size, shape, and any other discernible features.
Integration with Air Traffic Control (ATC): The information about UAPs should be integrated into existing air traffic control systems. This would allow for the dissemination of real-time information to pilots and ground control about potential UAPs in the vicinity of flight paths.
Automated Warning Systems: Developing automated systems to alert pilots and air traffic controllers when a UAP is detected in or near flight paths. This system should provide real-time updates on the UAP's location, movement patterns, and any potential risk it poses to aircraft.
Maneuvering Algorithms: Implementing algorithms that can provide recommended maneuvers or flight path adjustments to avoid UAPs. These algorithms should take into account the dynamics of both the aircraft and the UAP, as well as other environmental factors like air traffic, weather conditions, and terrain.
Pilot Training and Procedures: Pilots should be trained on how to respond to UAP encounters, including understanding the information provided by warning systems and executing recommended maneuvers. Standard operating procedures should be developed for UAP encounters to enhance safety.
Looking Forward
The release of NASA’s Unidentified Aerial Phenomena (UAP) Independent Study Team Final Report along with the recent unveiling of the U.S. Government’s All-domain Anomaly Resolution Office (AARO) website, the Department of Energy UAP website, and the availability of videos from U.S. Customs and Border Protection confirms that the government is gradually acknowledging the existence of yet to be explained phenomena operating in controlled airspace.
We do not definitively know their origins, intentions, or full capabilities. The potential for UAP to pose an imminent, significant threat is undeniable.
Navigating these unknown waters requires commitment, investment, and planning. AIAA UAP Integration and Outreach Committee subject matter experts are uniquely positioned to provide feedback and guidance to our stakeholders on the topics of detection, characterization, and evaluation of UAP. We look forward to working with industry, academic, and government partners on this challenging effort.
The AIAA UAP Integration and Outreach Committee will update this thought leadership article periodically.
