Elite Lanes: Evolutionary Generation of Realistic Small-Scale Road Networks

Artur Morys-Magiera; Marek Długosz; Paweł Skruch

cs.NEMar 21, 2026

Elite Lanes: Evolutionary Generation of Realistic Small-Scale Road Networks

Authors: Artur Morys-Magiera, Marek Długosz, Paweł Skruch

Organizations: AGH University of Krakow, Poland

Abstract

We present a comparative study of methods for generating realistic, constrained small- to medium-scale road networks with built-in redundancy. In this research, we evaluate the proposed Evolutionary Algorithm (EA) with connectivity and redundancy constraints against the Wave Function Collapse (WFC) method - commonly used in procedural terrain generation for games - and swarm algorithms: Particle Swarm (PSO) and Gray Wolf (GWO). Our focus is on producing realistic, redundant road networks suitable for vision, localization and navigation problems. We evaluate metrics: connectivity, cycles, intersections, dead ends, graph cut-edges while enforcing physical plausibility. We propose an EA and its extended version with elitism via MAP-Elites method. We detail the implementation, constraints, metrics and provide both visual and quantitative comparisons with baselines. Results highlight how fitness function design choices affect the structural characteristics of generated networks and highlight the impact of specific constraints in practical applications. Our contribution is a method for creating realistic synthetic datasets from sparse tile definitions derived from real-world data. We demonstrate a practical application by generating realistic maps using a laboratory-collected tileset from a Duckietown city model. Our approach performs coherent geometric transformations on metadata, in this work exemplified by semantic segmentation masks of the generated road networks.

Figures & tables

Figure 1: Elementary tile elements with associated semantic segmentation binary masks

Parameter	Value	Notes
Grid size	$12\times 12$	Tested across scales
Mutation rate	0.3

Table 2

Algorithm	Parameter	Value
PSO	Inertia weight	0.7
	$c_{1}$ (cognitive)	1.5
	$c_{2}$ (social)	1.5
	Generations	200
	Offspring	40
GWO	Generations	200

Table 3

Figure 2: Quantitative comparison of all models, according to metric definitions from section 2.2

	CO		DE		CY
	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
Method
EA	1.25	0.5	0.75	0.5	25	1.826
GWO	1	0	0.75	0.957	37.75	3.862
MAP-E	1.25	0.5	4.5	3.873	35.75	7.455
PSO	2.25	0.957	9.25	1.5	37	0.816

Table 1: Quantitative measures table 1 of 3, where: CO - connected components; DE - dangling (dead) ends; CY - cyclomatic complexity

	BV		ACV		CV
	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
Method
EA	0	0	81.25	22.66	1	0
GWO	0	0	341.8	38.06	1	0
MAP-E	0	0	305.2	95.66	1	0
PSO	0	0	334.5	19.16	1	0

Table 2: Quantitative measures table 2 of 3, where: BV - boundary violations; ACV - adjacent crossing violations scaled by connectivity; CV - coverage

	CR		SR		AT
	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
Method
EA	40.5	2.082	41.5	6.952	30.25	10.56
GWO	62.5	4.796	28.25	3.948	16.5	6.191
MAP-E	55.5	8.963	30.75	16.88	23.25	6.994
PSO	59.75	1.5	30.5	5.26	18.25	9.465

Table 3: Quantitative measures table 3 of 3, where: CR - crossings count; SR - quadratic consecutive straight roads length; AT - adjacent turns count

Figure 3: Example $14\times 14$ output generated by the WFC algorithm

Figure 4: Example $14\times 14$ output generated by the PSO algorithm

Figure 5: Example $14\times 14$ output generated by the GWO algorithm

Figure 6: Example $14\times 14$ output generated by the EA algorithm

Figure 7: Example $14\times 14$ output generated by the MAP-Elites algorithm

Graph Generation

Graph generation builds models and algorithms that create graphs with desired structures, properties, or distributions.

See all 67 papers

RoadWeaver: Large-Scale Lane-Level HD Map Generation from Scratch for Autonomous Driving Simulation
Aug 12, 2026Yueyuan Li, Zexi Chen, Weijie Xi +4HD Map Construction Procedural Content Generation
TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation
May 11, 2026Wilson Wongso, Lihuan Li, Arian Prabowo +4Trajectory Generation Synthetic Data Generation
University of New South Wales · Hong Kong University of Science and Technology (Guangzhou)
Looking Ahead and Back: Anticipatory Generative Modeling for Road-Network Trajectories
Oct 9, 2026Zhenhao Huang, Li Sun, Philip S. YuTrajectory Generation Constrained Generative Modeling
North China Electric Power University · Beijing University of Posts and Telecommunications · University of Illinois Chicago

Aug 12, 2026J/K move · Enter open · S save

RoadWeaver: Large-Scale Lane-Level HD Map Generation from Scratch for Autonomous Driving Simulation

Yueyuan Li, Zexi Chen, Weijie Xi, Mingyang Jiang+3

HD Map Construction Procedural Content Generation Autonomous Driving Simulation

Open paper AskPDF arXiv

Autonomous driving simulation requires diverse and scalable lane-level HD maps to support long-horizon evaluation across complex road networks. Existing approaches either rely on handcrafted or reconstructed real-world maps, which limits scalability, or generate only local road structures rather than complete HD maps. We present RoadWeaver, a coarse-to-fine framework for from-scratch generation of diverse, large-scale HD maps. RoadWeaver first synthesizes a global road layout, expands it into a connected road network, and then constructs lane-level geometry with topologically consistent lane connectivity. Experimental results show that RoadWeaver achieves a 99.8% reachability, a 10.7% dead-end ratio, and an endpoint alignment error of 0.24 m. Compared with SOTA generation methods, it reduces endpoint alignment error by 94.4% while generating complete HD maps in 1.39--3.50 s. The generated maps can be directly deployed in driving simulators, providing scalable simulation environments for future closed-loop evaluation of autonomous driving systems. The training code and an out-of-the-box implementation of RoadWeaver will be released upon acceptance.

Abstract

Graph Generation

Explore similar work

RoadWeaver: Large-Scale Lane-Level HD Map Generation from Scratch for Autonomous Driving Simulation

TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation

Looking Ahead and Back: Anticipatory Generative Modeling for Road-Network Trajectories