Should Drones Have a Speed Limit?

In this article we analyse the use of RemoteID (RID) working with Detect and Avoid (DaA) systems with the aim of enhancing airspace safety and make a conditional estimate of a “maximum speed” at which UA’s should be limited to.

Acronyms Used

ACAS Airborne Collision Avoidance System ADS-B Automatic Dependent Surveillance-Broadcast’ BVLOS Beyond Visual Line of Sight DaA Detect and Avoid DRIP Drone Remote Identification Protocol LoRa Long — Range Broadcasting TCAS Traffic-Alert and Collision Avoidance System UA Unmanned Aircraft UAS Unmanned Aircraft System UAV Unmanned Aircraft Vehicle

UA Adoption

As the skies become increasingly crowded, the importance of developing standardized protocols for remote identification, air traffic management, and collision avoidance has never been more important, underscoring the need for a coordinated approach to ensure manned and unmanned aircraft can co-exist.

UA Regulations

The rapid adoption of UAS presents challenges in terms of privacy, security, and airspacemanagement. Regulatory authorities like the Federal Aviation Authority (FAA) in the United States have mandated remote identification (Remote ID) for drones to enable real-time tracking and management of airspace. In EU the ASD-STAN, the European body for Aerospace standards, are responsible for the regulations. The prEN 4709–002 requires Remote ID for drones that are flying below 120 meters and weigh more than 250 grams.

Remote ID

Remote ID functions as a digital “license plate” for drones, providing important information such as the drone’s ID, location, altitude, and launch location. This information is important for airspace awareness among all users and authorities, for increased safety and security. The technology standards for Remote ID have been developed to ensure a universal broadcasting method that is available and understandable across different platforms and devices.

ASTM

The ASTM F3411–19 Standard Specification for Remote ID and Tracking establishes guidelines for broadcasting remote identification information via various communication methods including Bluetooth and Wi-Fi. This standard is designed to ensure drones can be identified remotely, promoting safety and accountability in airspace. The emphasis is on compatibility and minimal implementation burden, highlighting mandatory and optional message fields and formats to facilitate broad acceptance and integration.

ASD-STAN

The ASD-STAN prEN 4709–002 standard focuses on Direct Remote Identification (DRI), specifying the digital identification requirements for UAS in certain classes. It addresses the operational concept, European regulation and standardization, and outlines communication technologies and protocols. Key features include broadcasting operator ID and drone information using Bluetooth and Wi-Fi, with emphasis on visibility and accessibility to enforcement authorities and the public.

DRIP

The Drone Remote Identification Protocol (DRIP). [DRIP] is a protocol focused on addressing the identification and tracking of unmanned aircraft systems at an internet protocol level, going beyond the capabilities offered by ASTM and ASD-STAN standards. DRIP seeks to extend Remote ID functionalities to support advanced operational needs such as integration with manned aircraft systems, enhanced privacy features, and secure information sharing across networked environments. While drip offers a more secure solution it also comes with latency overhead to encode and decode messages. While the DRIP initiative is still under development, it represents a significant step towards achieving global standardization for Remote ID.

DAA

Detect and Avoid (DAA) technology ensures the safe operation of UAS within shared airspace. DAA systems are developed to enable drones to autonomously detect potential collision threats and navigate away to prevent accidents. This technology is critical for developing drone applications, especially for Beyond Visual Line of Sight (BVLOS) operations where direct visual monitoring is not possible. DAA is key for integrating drones safely into the airspace, alongside manned aircraft, enhancing both efficiency and safety. DAA algorithms are mathematical models that run on the computer of the drone or connected hardware to detect these collisions. These algorithms can use data like location data of drones in its perimeter and data from Lidar or Radar sensors.

ACAS

The Airborne Collision Avoidance System (ACAS) is a DAA implementation. Initially developed for manned aviation as TCAS (Traffic Collision Avoidance System), ACAS has evolved to address the unique challenges posed by the integration of Unmanned Aerial Systems (UAS). ACAS systems, including variations such as ACAS Xu for unmanned aircraft, ACAS sXu for drones and many others, utilize advanced algorithms to analyze the positional data of surrounding aircraft. This data allows the system to predict potential conflict scenarios. When a threat is detected, ACAS dynamically generates avoidance advisories, guiding pilots or autonomous drone systems to take action.

Integrating With Manned Aircraft Systems

TCAS

Traffic Collision Avoidance System (TCAS) monitors the airspace around an aircraft for other aircraft equipped with a corresponding active transponder, independent of air traffic control, and warns pilots of the presence of other transponder-equipped aircraft which may present a threat of mid-air collision (MAC). TCAS is mandated for use in many types of passenger- carrying aircraft under certain regulations, with varying requirements depending on the region and the aircraft’s operation type. For UAS integration, the challenge lies in ensuring drones can operate safely in proximity to manned aircraft that rely on TCAS for collision avoidance. Adapting UAS operations to be compatible with TCAS involves ensuring drones are detectable by TCAS systems and can respond appropriately. This could involve the development and implementation of protocols that enable drones to mimic the behavior of manned aircraft in response to TCAS advisories, either by adjusting their flight path autonomously or by providing alerts to remote pilots for manual intervention.

ADSB

Automatic Dependent Surveillance-Broadcast (ADS-B) is a surveillance technology in which an aircraft determines its position via satellite navigation and periodically broadcasts it, enabling it to be tracked. The information can be received by air traffic control ground stations as a replacement for secondary radar as it provides more accurate tracking. Unlike radar, ADS-B works at all altitudes, is more accurate, and provides several safety-critical services to pilots and air traffic controllers. Integrating drones into airspace with ADS-B involves equipping drones with ADS-B transmitters and, potentially, receivers to ensure they can both broadcast their positions and receive the positions of nearby aircraft. This integration makes the situational awareness for both manned aircraft and drones possible, enhancing safety. The challenge lies in ensuring the ADS-B system can handle the increased number of transmitters without overwhelming the system and that drone operations do not interfere with the ADS-B signals from manned aircraft. Potential solutions include developing UAS-specific ADS-B frequencies or protocols, or leveraging ADS-B data for improved traffic management and collision avoidance algorithms for drones.

The Theory

Which Protocol?

Latency

Both Bluetooth and Wi-Fi utilize radio waves for wireless data transmission, with these waves travelling at speeds close to that of light. Therefore, the velocity speed of radio waves does not impact latency. Instead, latency can be attributed to factors such as encod- ing/decoding processes, protocol efficiency, and interference management, among others. In the comparative performance study of ESP-now, Bluetooth 4.2, and Wi-Fi, there were four kinds of trials to measure latency: 1, 10, 50, and 100 bytes. The study showed that increased packet size increases latency in ESP-NOW and Bluetooth, where 1 byte took 6.0 ms for Bluetooth, 10 bytes 6.2 ms, 50 bytes 6.4 ms, and finally 100 bytes 8.5 ms. In comparison, Wi-Fi had an almost fixed latency of about 3.4 ms for all of the packet size, indicating its suitability for applications where timely data transmission is critical. The fact that packet size has minimal impact on latency ensures consistent performance across a range of data sizes. In order for DRIP to be fully viable its broadcasted in 20 bytes, limited with respect to bandwidth, battery life and processing power. This means that DRIP would only cause a fixed latency of about 6.2–6.4 ms, which would not have a significant effect on the reaction time.

Range

Bluetooth shows a marked improvement in range when switching from a built-in to an external antenna, jumping from 15 meters to 25 meters. This increase highlights Blue- tooth’s adaptability and the potential for range extension through hardware modifications, which can be particularly beneficial in extending connectivity for personal area networks. This is also supported by another research in”Bluetooth 5: A Concrete Step Forward toward the IoT” where it supports the range claim for Bluetooth 4.x, stating it has a range between 50 and 100 meters outdoors and unobstructed, which is reduced to 10/20 meters in indoor environments. The same study showed that Bluetooth 5 aims to significantly increase the range of Bluetooth devices. In the worst case, this range should be 200 meters outdoors and about 40 meters indoors. This range is exceeded in Bluetooth 5.2 which introduces the Long Range mode, capable of achieving up towards 1 km in outdoor range, which aligns with the expected range of ASD-STAN expected range of Bluetooth 5. In comparison, their expected range of Wi-Fi NAN is upwards 2 km in ideal conditions such as an unobstructed line of sight, optimal antenna implementation, maximum allowed TX power, and without any interfering elements.

Wi-Fi’s effective indoor range is significant for practical applications, with a base range of 84 meters using the built-in antenna, which slightly improves to 88 meters with an external antenna. This relatively modest increase with an external antenna suggests that while enhancements are possible, environmental factors and interference may play a larger role in range limitation than antenna type alone.

Data Transmission

With a transmission speed of 938 Kbps and an Maximum Transmission Unit (MTU) of 251 Bytes, Bluetooth is well-suited for applications that do not demand the high bandwidth provided by Wi-Fi. Its design for lower data rate applications, such as peripheral connectivity or simple data exchange between devices, reflects its optimal use in scenarios not requiring intensive data transmission. Wi-Fi stands out for its superior data transmission speed, offering up to 2048 Kbps, making it the fastest among the protocols analyzed. Its high MTU of 1460 Bytes allows for efficient data packet transfers, catering to applications requiring high bandwidth and rapid data exchange. The protocol’s design to minimize packet loss through specific transmission delays further underscores its robustness for demanding applications.

Power

Bluetooth excels in energy efficiency, with significantly lower power consumption than Wi-Fi. This advantage is critical for battery-powered devices, such as drones, where extending operational life is crucial. Its efficient power usage, particularly in standby or low-data scenarios, supports its widespread adoption in the consumer electronics sector. With a power consumption of approximately 650 mW, Wi-Fi is more energy-intensive compared to Bluetooth. This characteristic makes it less suitable for applications where energy efficiency is paramount, particularly in portable or battery-operated devices. However, its lower power usage when acting solely as a receiver offers some flexibility.

Signal Strength

In a study made in 2020 conducted an experiment where the signal behavior of Long — Range Broadcasting (LoRa) technology was studied across urban, suburban, and forest environments. It was found that signal stability greatly depends on the environment, being more stable in suburban areas compared to high-density urban areas. In forests, the primary challenge is the reflection of signals off vegetation, whereas, in urban environments, reflections off buildings and vehicles predominate. The forest environment is particularly challenging for radio wave propagation due to vegetation, climate variations, and Beyond Visual Line of Sight (BVLOS) conditions. Urban scenarios, in contrast, have their own set of challenges, including the variability introduced by buildings, vehicles, and other elements that creates a highly dynamic environment.

Measuring Distance

Remote ID system broadcasts a drone’s geographical point using latitude and longitude coordinates. In order to calculate the distance between different points either a simplified approach can be taken with or without or the Earth’s spherical shape in consideration. Haversine Formula is a an equation commonly used in navigating and calculating distance on a sphere.

Flat Earth approximation is a simpler, but albeit less accurate method for estimating the distance between two points on Earth by assuming a flat surface. For shorter distances (less than 20 kilometres) it can be assumed to be accurate enough.

UA Capabilities

Quadcopter

Quadcopters maneuver style is characterized by the ability to hover, fly in any direction, and perform agile turns. This agility is largely due to the independent control of the speed of each rotor. Regarding deceleration, quadcopters can decelerate and stop mid-air relatively quickly due to their ability to reverse thrust almost instantaneously. The exact deceleration rate can vary significantly between models depending on size, but we estimate it to be around 5.5m/s2 for a larger model.

Wing Based

Wing-based aircraft, for example fixed-wing drones and manned airplanes, rely on forward motion to generate lift over their wings. An important aspect of their maneuver capabilities is the stall speed, which is the minimum speed at which an aircraft must travel to maintain adequate airflow over the wings for lift. If the aircraft’s speed drops below this threshold, it can lose lift and potentially enter a stall, where it begins to descend or fall uncontrollably. The stall speed varies depending on the aircraft’s design, weight, and configuration, such as the deployment of flaps. In a study on an experimental flying wing UAV , the stall speed was discussed in the context of flight testing. The UAV’s stall speed was approximately 11m/s during one of the flight tests, which was close to the estimated stall speed of 10.4m/s given the UAV’s actual weight. With a manned airplane like a Boeing 737 we estimate the stall speed to be much higher, around 100m/s to generate the much more lift force that is needed. This is why most DAA methods for manned airplanes uses vertical avoidance.

Modelling Parameters

The easiest essential parameters to take into account for modelling the maximum safe speed for UAS are:

• Broadcast Range • Latency • Deceleration

Modelling

We use a formula to calculate the velocity based on initial velocity, acceleration, distance, and reaction time can be used. In our case the reaction time would be interpreted as a latency for the drone to get the broadcast signal and the overhead of validating it. This could be various calculations our authentication solution like DRIP where the drone needs a response from the ground station whether the intruding drone is valid or not.

Simulation

In the simulations, the goal was to observe the effects of latency, broadcast range, and velocity on the drones’ ability to detect each other before a collision. The benefit of using a simulated model in addition to the mathematical, is that we could actually see how the different combinations of the parameters actually affected one another. We assume these parameters can vary depending on various conditions so we will also look at scenarios where the broadcast range and latency are really bad, arguably even unrealistic. The overall scenario involved two drones following predetermined paths towards each other. In each scenario, they were on a collision course, but the initial conditions could vary. Each drone could have different set initial velocities, broadcast ranges, and latencies. The starting velocities ranged from 10 to 100 m/s, the broadcast ranges from 10 to 1000 m, and the latencies varied between 0.1 and 5 seconds. For this study, 3000 different scenarios were simulated. In each scenario, we recorded whether a collision was avoided through detection or if it occurred before detection. Additionally, we noted the latency, velocity of drone 1, velocity of drone 2 speed, and their individual broadcast ranges for each scenario. The aim was to see if there was a way of identifying a minimum threshold for each parameter to minimize the risk of collision.

Analysis

Analyzing the simulation results allows us to identify specific conditions and parameters that led to collisions. Notably, collisions occurred when the combined speed approached 75 m/s, under challenging conditions with a short broadcast range and higher latency. Within the confines of this simulation, running 3000 simulations and considering only two parameters in head to head scenarios there were no instances of collisions at speeds below this threshold, suggesting it could represent a maximum. However, this assertion is doubtful, given the simulation’s limited scope in parameters and scenarios. A lower threshold for maximum safe speed is plausible when a broader spectrum of external conditions is taking into account.

So, our determination is that 75 m/s should be the absolute maximum flight speed for a UA operating in an urban environment.

Cheska Gal for DeFli Networks