Multilayer Evaluation of Vertiport Departure Directions Under Urban Obstacles and Wind

arXiv:2610.21324 · eess.SY, cs.SY · Submitted 2026-10-15 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.

Rosa: Today's paper: "Multilayer Evaluation of Vertiport Departure Directions Under Urban Obstacles and Wind".

Dev: The gist The framework evaluates every departure azimuth and identifies usable and preferred departure directions for each climb capability.

Rosa: First, who's behind it and why it matters.

Title and authors: Rosa: We've seen how the paper sets up this evaluation stage by stage, focusing on different climb angles. Now let's look at what the authors actually summarize in this paper, "Multilayer Evaluation of Vertiport Departure Directions Under Urban Obstacles and Wind."

Dev: They summarize that the framework combines two main assessments: a vertical obstacle assessment score, Sobs which depends on theta and g, with a lateral assessment for obstacles. The lateral part acts as a strict filter first to make sure you're in the protected corridor defined by the required navigation performance or RNP.

Taro: So it starts by saying if you're not in the corridor defined by the RNP, you don't even bother looking at how much room you have vertically, which makes sense for safety constraints first.

Rosa: Right. Then they perform that vertical assessment of the obstacle clearance margin for all the remaining azimuths, and they use a multiplicative formulation to combine these scores into one obstacle clearance score Sobs for each azimuth theta in each climb angle layer g.

Dev: The paper then breaks down how this score is calculated using several metrics, specifically the vertical obstacle assessment SV ER which compares the obstacle height HOBS with the OCS and climb path geometry, and the lateral assessment SHOR which uses a parameterized power law normalization to control how quickly that score decays near the OCS boundary.

Taro: So we’ve got these specific metrics like SV ER and SHOR, but then they combine them with wind operability scores too, which means it’s not just about static obstacles anymore; the wind plays a big role in whether that direction is actually usable.

Rosa: Right. And then there's the Wind Operability Score which is quantified by multiplying the prevailing wind score and the mean crosswind operability, where that prevailing wind score Spw expresses how often you’d fly into a headwind relative to the most favorable azimuth, and Sac is just the average of instantaneous crosswind scores over a ten year period.

Dev: So when you put it all together, this entire structure is designed to evaluate every departure azimuth over the whole directional domain in a dense urban obstacle environment while quantifying that obstacle clearance margin in each direction.

The paper's summary: Rosa: Now for what’s interesting, the paper doesn't just stop at describing the setup; they propose some real improvements to this evaluation framework that we should look at.

Dev: They suggest that by using this multilayer approach, you get a ranked set of usable directions rather than just a single preferred direction, which gives operators more flexibility. This means instead of one answer, they give you a list.

Taro: So the improvement is moving from finding *one* best path to generating a candidate set. That lets you check that whole set against the current wind at the operational stage and pick the one with a smaller crosswind component if needed.

Rosa: Right. And they suggest that this framework can be used by traffic management models where you treat the choice of procedure as a decision variable and its usability as a constraint, which is huge for automated systems.

Dev: They also show how you can use the output to determine the preferred departure direction jointly by considering operational priorities, the climb profile available to the aircraft, and how obstacles and prevailing winds are arranged around the site.

Taro: So it moves beyond just calculating a score; it gives you a way to make decisions that balance safety against efficiency in real-time based on what’s happening on the ground.

Rosa: They also point out that this framework is computationally efficient enough to run at many candidate vertiport sites before detailed design, so it’s useful early in the planning process.

The paper's improvements: Dev: So to wrap up, the authors are showing how this multilayer evaluation of vertiport departure directions under urban obstacles and wind provides a way to secure operational flexibility when different eVTOL aircraft share one vertiport.

Rosa: They conclude that no single departure direction is preferred at a site; instead, the choice follows from operational priorities, the climb profile available to the aircraft, and the arrangement of obstacles and winds around it.

Taro: From my side, I think this is important because it gives a concrete procedure for how traffic management models can treat procedure selection as a decision variable and its usability as a constraint.

Dev: Right. And they emphasize that this framework can be retained as procedures available to those models, where the choice of procedure is the decision variable and the usability of each one is a constraint.

Rosa: So, "Multilayer Evaluation of Vertiport Departure Directions Under Urban Obstacles and Wind," it shows how you can integrate urban obstacle conditions and wind operability within a single scoring structure to evaluate every azimuth for different climb angle layers.

Taro: That gives operators a ranked set of usable directions that they can then check against real-time wind data to choose the one with the smaller crosswind component.

Dev: It’s a framework that integrates these factors, and it's computationally efficient enough to be run at many candidate vertiport sites before detailed design.

Conclusion: Rosa: So we've looked at how this paper, "Multilayer Evaluation of Vertiport Departure Directions Under Urban Obstacles and Wind," evaluates every single departure azimuth and figures out which ones are usable for different climb capabilities.

Dev: Exactly, it’s not just about finding one good path; it’s generating a ranked set of directions that lets you choose based on the current conditions.

Taro: I think what stands out is how they combine the vertical obstacle assessment with the lateral assessment—that first filter based on RNP constraints.

Rosa: Yeah, and then they layer in wind operability, using that product of the prevailing wind score and the mean crosswind operability to get a total score.

Dev: The numbers are interesting, like how at one site, only eight point nine percent of the azimuth space was usable for a five-degree climb layer. That shows how much the environment dictates what’s possible.

Taro: It really highlights that the same climb performance can secure very different directional resources depending on where you are—like Suseo station versus Jamsil heliport in Seoul.

Rosa: And they show that the preferred departure direction isn't fixed; it depends jointly on operational priorities, the aircraft's climb profile, and everything around it.

Dev: So this framework can be retained as procedures for traffic management models, letting them treat the choice of procedure as a decision variable and its usability as a constraint.

Taro: It’s efficient enough to run at many sites before detailed design starts, which makes it useful early in the planning process.

Rosa: Overall, this study gives us a concrete way to integrate urban obstacle conditions and wind operability into one scoring structure for evaluating every azimuth across all climb angle layers for eVTOL vertiports.

Dev: It’s a practical tool for figuring out how to manage flexibility when different aircraft types share the same landing spot.

Taro: We can see how this approach feeds directly into conflict-aware procedure selection, letting an operator pick a direction with a smaller crosswind component when the wind gets rough.

Rosa: Yeah, it moves us past finding a single runway style path and toward giving operators a whole set of candidate directions to work with.

Dev: Next up, we’re looking at how this kind of dynamic resource allocation works in the context of larger multi-robot systems and state governance.

Seunghyeop Jeong, Max Z. Li, Kyowon Song

Kookmin University · University of Michigan

eess.SY, cs.SY

Submitted: 2026-10-15

Updated: 2026-10-15

Comments: 37 pages, 13 figures, 5 tables

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 76/100

The gist: The gist The framework evaluates every departure azimuth and identifies usable and preferred departure directions for each climb capability.

Key concepts

Obstacle Clearance Score (Sobs)
This score combines vertical and lateral assessments of nearby buildings. It first acts as a strict filter to remove directions where an obstacle violates required navigation performance rules, then assesses the remaining directions based on how much physical space is left around the aircraft.
Vertical Obstacle Assessment (SV ER)
This measures the safety margin below the required navigation performance height. It compares the actual obstacle height against a calculated operational clearance surface. A positive score means there is a secured margin below this critical height, indicating safe vertical separation.
Wind Operability Score
This quantifies how favorable a specific departure azimuth is regarding wind conditions over ten years. It combines the prevailing wind score (how often the route hits a headwind) with the mean crosswind operability, giving a holistic view of directional flight feasibility.
Climb Angle Layers
These layers categorize aircraft based on their different climb performance capabilities. By evaluating directions separately for each layer, the framework accounts for how different aircraft types can utilize or avoid specific departure paths based on their ascent characteristics.

Summary

The gist The framework evaluates every departure azimuth and identifies usable and preferred departure directions for each climb capability.

How it works

The proposed framework evaluates vertiport departure directions over the entire directional domain in a dense urban obstacle environment and quantifies the obstacle clearance margin in each direction. The evaluation follows a staged structure that progressively narrows the candidates from geometric usability to directional operability, and is performed independently for each climb angle layer.

The core of the framework is the combined obstacle clearance score Sobs (theta, g) of azimuth theta in climb angle layer g using a multiplicative formulation The lateral assessment acts first as a strict usability filter, excluding azimuths that violate the protected corridor defined by required navigation performance (RNP) The vertical assessment of the obstacle clearance margin is then performed for the remaining azimuths.

Key Evaluation Components

The framework utilizes several key metrics to quantify suitability:

- Vertical Obstacle Assessment:

Vertical assessment evaluates how much margin remains even when no penetration occurs. The vertical score SV ER is determined by comparing the obstacle height HOBS with the OCS and climb path geometry. A zero or positive value means that an obstacle reaches or exceeds the OCS height at its location, and a negative value indicates a secured margin below the OCS.

- Lateral Obstacle Assessment:

The lateral assessment considers obstacles that lie within the buffer zone on either side of the centerline and whose height reaches the protection surface. The normalized lateral score SHOR is defined using a parameterized power law normalization where exponent p controls how quickly the score decays near the OCS boundary.

- Wind Operability Score:

Wind operability is quantified by the product of the prevailing wind score and the mean crosswind operability The prevailing wind score Spw expresses how often a departure along that azimuth would be flown into a headwind, relative to the most favorable azimuth. The mean crosswind operability Sac is defined as the average of instantaneous crosswind scores over the ten year period

Application and Findings

The study applied this framework to two candidate sites in Seoul: Suseo station and the Jamsil heliport. The results showed that the same climb performance can secure very different directional resources at different sites. For instance, at Case 1 (Suseo station), the 5◦ layer yielded only 8.9% of the azimuth space as usable.

Operational Implications

The final output is a ranked set of usable directions rather than a single direction. This allows for operational flexibility where a candidate set evaluated in advance can instead be checked against the current wind at the operational stage, and a usable direction with a smaller crosswind component chosen. The results support policy decisions by showing how the preferred departure direction is determined jointly by operational priorities, the climb profile available to the aircraft, and the arrangement of obstacles and prevailing winds around the site. This framework can be retained as procedures available to a traffic management model with the choice of procedure as a decision variable and the usability of each procedure as a constraint. The framework is computationally efficient enough to be run at many candidate vertiport sites before detailed design.

Conclusion

This study addressed how operational flexibility in the departure direction can be secured when eVTOL aircraft with different climb and crosswind capabilities share a single vertiport. It proposed a framework that integrates urban obstacle conditions and wind operability within a single scoring structure, evaluates the full 360◦ of azimuth for each climb angle layer, and derives a ranked set of usable directions for each layer. The study concludes that no single departure direction is preferred at a site; the choice follows from operational priorities, the climb profile available to the aircraft, and the arrangement of obstacles and winds around it. The study is limited by using representative scenarios for climb angles and crosswind limits. Future work should incorporate type specific performance as it emerges from certification and flight testing. The funding sources include support from the National Research Foundation of Korea grant funded by the Korea government (Ministry of Science and ICT) and the Korea Agency for Infrastructure Technology Advancement grant funded by the Ministry of Land, Infrastructure and Transport.

Improvements for AI systems

  1. textbf Directional Usability Framework Integration in Planning Systems: The improved AI system can evaluate every departure azimuth and identify usable and preferred departure directions for each climb capability, moving beyond a single runway style path. This allows planning systems to generate a set of candidate directions matched to the obstacles and wind of the site and the climb capability of the aircraft.

  2. textbf Multilayer Fleet Performance Modeling: The system can utilize three climb angle layers reflect differences in climb capability among aircraft to determine operational suitability, specifically showing how the number of usable azimuths varies with the allowable climb profile. This enables dynamic allocation of vertiport resources based on the specific eVTOL type arriving.

  3. textbf Dynamic Operational Prioritization via Weighting Scenarios: The system can incorporate four weighting scenarios between obstacles and wind reveal how operational priority affects the preferred departure direction, allowing decision-making under uncertainty. This lets operators select procedures by prioritizing safety (Obstacle dominant) or efficiency (Wind prioritized).

  4. textbf Candidate Set Generation for Traffic Management: The system can generate ranked candidate sets P(k)g,Vc for each layer and crosswind limit, which are then sorted by the integrated score S(k)total. This output serves as a direct input to higher level traffic management models, allowing them to treat the choice of procedure as a decision variable.

  5. textbf Conflict-Aware Procedure Selection: The system can check these pre-screened candidates against real-time wind data, allowing an operator to select a usable direction with a smaller crosswind component chosen from the available set. This capability directly addresses the need to choose an alternative procedure when crosswind component approaching the allowable limit.

Related papers