Summary
Background
Urban airspace is a unique environment where highly obstructed (complex topology) and sparsely structured (simple topology) regions coexist. The presence of static obstacles and the required separation from those obstacles creates particular spatial structures - narrow passages, islands, and funnels - throughout the airspace.
Objective
This study proposes a computational geometric approach to represent urban airspace as a compact and informative graph structure that summarizes spatial connectivity in the horizontal and vertical dimension.
Contribution
Whether a navigable space is highly obstructed or sparsely structured has important implications for airspace management, since not only the throughput capacity of airspace [1-3] but also the complexity of path planning [4-8] are highly dependent on the topology of the naviable space. One can utilize the proposed graph structure to derive the sequence of airspace segments that Unmanned Aircraft (UA) can use.
The problem of allocating airspace volumes to each vehicles can be decomposed into separate problems at the individual segment level, which could reduce the overall complexity of flight planning in the low-altitude airspace.
Related publication
Cho, J., & Yoon, Y. (2021). Extracting the topology of urban airspace through graph abstraction. Accepted to Transportation Research Part C: Emerging Technologies, 127, 103116
Cho, J., & Yoon, Y. (2020). Segmentation of Low-altitude Airspace in Highly Constrained Environments, International Conference on Research in Air Transportation (ICRAT 2020). [pdf]
Method
Decomposing the airspace into segments and passages
We call a portion of airspace segment if it allows at most n vehicles having containment radius r to simultaneously pass through without intruding unavailable space.
The concept of a segment is adopted from the topological concept called pocket, which is the term that defines imperfect holes in computational geometry [9].
The configuration of a segment is determined by a width parameter α, where α=n∙r is the radius of a disc that can fit in the segment. Each segment is connected to the outside only via relatively narrow passages, where a passage is a part of airspace that n vehicles cannot enter or exit at once. Surrounded by passages, each segment can be viewed as a region having limited accessibility from neighboring regions.
This figure shows an illustrative example of one segment (in green) linked to the outer space through three passages (in orange) with respect to a fixed containment radius r and a width parameter α=3∙r.
Obstacle space is colored in black, and unavailable space are shaded in grey. Segment and passage are indicated by green and orange boxes, respectively
Graphical abstract of segment configuration
The figure on the right shows the result of increasing α in segment configurations. The polytope in green represents available airspace where n vehicle with containment radius r can enter its space without touching any obstacles.
Increasing α=0 to α=3∙r results in splitting airspace into three disjoint segments. By tracking how the segment connectivity evolves with parameter α, one can construct a segment graph that encodes the adjacency between segments and passages.
Application of the alpha shape filtration to a hypothetical environment; (a) segment configuration and (b) corresponding segment graph
Results
The figure on the right shows segment configurations and corresponding segment graphs at altitude of 60m and 80m.
There are a greater number of segments (=40) at ℎ=60. One segment is connected to only about two segments on average, indicating low accessibility within this airspace layer.
Segments at lower altitudes are generally smaller in size and are scattered throughout the airspace, as the lower airspace is more heavily constrained by geospatial elements.
In contrast to the lower airspace at ℎ=60, the upper airspace at ℎ=80 has fewer segments (=18) and a higher average degree (=3.1). As the altitude increases, new segments are formed, lower altitude segments merge into larger ones, and lower altitude passages also develop into segments.
(Left) Available and unavailable airspace (Middle) airspace segments and passages (Right) segment graph at 80 m with respect to keep-out geofence of 30 m and keep-in geofence of 10 m
The figure on the right shows the Reeb graph of airspace segments in the case study area.
The Reeb graph structure clearly shows at which altitude individual segments are formed and connected to other segments. The formation of new segments is most frequent at 50≤h≤60, while the merging of lower-altitude segments to higher-altitude segments is at the highest level at 70≤h≤90.
It also shows that a high proportion of narrow air passages exists below 70 m and that larger open segments with improved accessibility exist above 110 m.
Airspace above 130m, which is mostly unobstructed, is rarely affected by the alpha shape filtration, indicating the capability of the proposed method to distinguish topologically complex and simple environment.
The Reeb graph of airspace segments in the study area
Conclusion and future directions
Conclusion
Urban airspace is a unique environment where highly obstructed (complex topology) and sparsely structured (simple topology) regions coexist. The formation of several spatial structures, such as narrow tunnels, islands, and bottleneck points, creates difficulties in locating airspace volumes to accommodate large-scale aerial activities. This study divides the entire airspace into multiple segments connected via relatively narrow passages. Low-altitude urban airspace is represented as a graph structure that summarizes the segment connectivity in the horizontal and vertical dimension.
Limitation and further study
The results of the case study are limited to the area under analysis, but the proposed techniques are applicable to other airspaces and can also be extended to larger regions. One way of applying the method to larger regions is to compute segment configurations in several adjacent regions of interest and merge the configurations of adjacent regions of interest (see the figure below)
While this paper only shows case study results given one set of parameters, the methodology is designed in a way that produces outputs with any user-specific parameter combinations. One may adaptively define and construct segments and passages depending on the underlying shape of a given airspace rather than using a fixed value. Investigating the effect of varying separation distances and width parameters on airspace configurations is another research topic, especially to find the volume of airspace that remains persistently available in regard to changing parameters.
Illustrative example of extending the proposed decomposition method to a larger region
References
Mitchell, J., Polishchuk, V., & Krozel, J. (2006). Airspace throughput analysis considering stochastic weather. In AIAA Guidance, Navigation, and Control Conference and Exhibit (p. 6770).
Krozel, J., Mitchell, J. S. B., Polishchuk, V., & Prete, J. (2007). Capacity estimation for airspaces with convective weather constraints. Collection of Technical Papers - AIAA Guidance, Navigation, and Control Conference, 2, 1518–1532. https://doi.org/10.2514/6.2007-6451
Krozel, J., Mitchell, J. S. B., Polishchuk, V., & Prete, J. (2007). Maximum Flow Rates for Capacity Estimation in Level Flight with Convective Weather Constraints. Air Traffic Control Quarterly, 15(3), 209–238. https://doi.org/10.2514/atcq.15.3.209
Wang, K. H. C., & Botea, A. (2011). MAPP: A scalable multi-agent path planning algorithm with tractability and completeness guarantees. Journal of Artificial Intelligence Research, 42, 55–90. https://doi.org/10.1613/jair.3370
Sharon, G., Stern, R., Felner, A., & Sturtevant, N. (2012). Meta-agent conflict-based search for optimal multi-agent path finding. Proceedings of the 5th Annual Symposium on Combinatorial Search, SoCS 2012, 97–104.
Sharon, G., Stern, R., Felner, A., & Sturtevant, N. R. (2015). Conflict-based search for optimal multi-agent pathfinding. Artificial Intelligence, 219, 40-66.
Boyarski, E., Felner, A., Stern, R., Sharon, G., Tolpin, D., Betzalel, O., & Shimony, E. (2015). ICBS: improved conflict-based search algorithm for multi-agent pathfinding. In Twenty-Fourth International Joint Conference on Artificial Intelligence.
Ryan, M. R. K. (2008). Exploiting subgraph structure in multi-robot path planning. Journal of Artificial Intelligence Research, 31, 497–542. https://doi.org/10.1613/jair.2408
Edelsbrunner, H., Facello, M., & Liang, J. (1998). On the definition and the construction of pockets in macromolecules. Discrete Applied Mathematics, 88(1–3), 83–102. https://doi.org/10.1016/S0166-218X(98)00067-5