Reasonable Motion: A General ASP Foundation for Environment Constrained Movement Trajectory Computation

Jakob Suchan; Julius Monsen; Lars Karlsson; Mehul Bhatt

arxiv: 2606.25626 · v1 · pith:NMTQMD2Unew · submitted 2026-06-24 · 💻 cs.AI · cs.HC· cs.RO

Reasonable Motion: A General ASP Foundation for Environment Constrained Movement Trajectory Computation

Julius Monsen , Jakob Suchan , Mehul Bhatt , Lars Karlsson This is my paper

Pith reviewed 2026-06-25 20:40 UTC · model grok-4.3

classification 💻 cs.AI cs.HCcs.RO

keywords answer set programmingtrajectory computationmotion planningautonomous drivingstable modelshybrid reasoningenvironment constraintsgraph traversal

0 comments

The pith

Answer set programming encodes geometric and domain constraints as rules whose stable models represent admissible motion trajectories.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a hybrid quantitative-qualitative method that uses answer set programming to compute constrained branching trajectories for moving objects. It encodes an environment as a graph and treats stable models of the resulting program as distinct trajectory modes that satisfy both spatial geometry and higher-level norms. Each mode carries traceable information about event sequences, topology, and rules, which supports interpretability not available from purely learned systems. The authors demonstrate the approach on real-world autonomous driving data to show applicability across varied motion patterns.

Core claim

The method performs constrained traversal of an environment graph, enumerating geometrically admissible motion behaviours as stable models, each constituting a distinct trajectory mode characterised by both domain-dependent and independent factors such as derived event sequence, map topology, and domain norms. The hybrid trajectory computation method is generally applicable across motion characteristics typically encountered in diverse dynamic domains with moving objects.

What carries the argument

Encoding of geometric constraints, map topology, and domain norms as ASP rules so that stable models correspond to admissible trajectory modes during graph traversal.

If this is right

Trajectories remain traceable to the specific stable model that generated them.
The same encoding framework applies to motion patterns in domains beyond autonomous driving.
Branching modes arise naturally from the enumeration of stable models.
The approach supplies verifiable compliance with both geometry and domain norms.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The rule-based encoding could be extended to verify safety invariants before execution.
Dynamic updates to the ASP program might support online replanning when new sensor information arrives.
Multi-object interactions could be added by introducing additional rules that couple separate trajectory programs.

Load-bearing premise

Domain norms, geometric constraints, and map topology can be effectively encoded as ASP rules such that stable models correspond to geometrically admissible and norm-compliant trajectory modes.

What would settle it

A concrete case in the Argoverse 2 dataset where an observed valid trajectory has no corresponding stable model under the encoding, or where a generated stable model yields a geometrically invalid path.

Figures

Figures reproduced from arXiv: 2606.25626 by Jakob Suchan, Julius Monsen, Lars Karlsson, Mehul Bhatt.

**Figure 1.** Figure 1: Hybrid Qualitative-Quantitative Motion Trajectory Computing Method: Overview of the proposed pipeline. Left: ASP grounding and solving yield stable models, each representing a path annotated trajectory influencing events. Right: stable models are grouped and geometrical features constructed such that paths can be followed continuously according to a domain-specific motion model and yield a trajectory traj… view at source ↗

**Figure 2.** Figure 2: Reachability over G; Example: • current agent position; • potential start node; • reachable nodes. Solid arrows denote successor transitions; dashed arrows denote twohop lateral relations (as used in Sec. 3.1). T ⊂ T over which trajectory modes are computed. The ASP program encoding the scene is then: Π = G ∪ O ∪ T, where G = (N , E) is a typed directed graph whose nodes and edges are grounded as facts, w… view at source ↗

**Figure 3.** Figure 3: Discrete Trajectory Modes. Each mode is highlighted by a path between the lane segments and their representative points •. • denotes the position of the agent. INT[S/L/R] denotes an intersection event where the agent goes straight, left or right, and LC[L/R] a left or right lane change. representative path Ci , lane segments of each model are stitched into a continuous polyline. At junctions, a cubic Bézi… view at source ↗

**Figure 4.** Figure 4: ASP statistics. Top: distributions of solving time (left) and grounding time(right) on a log scale; dashed lines indicate medians. Bottom: total runtime (left) and number of conflicts (right) as a function of the number of ground rules [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗

**Figure 5.** Figure 5: Predicted Trajectories for Three Argoverse 2 Scenarios. Each coloured path corresponds to one trajectory mode, labelled by its event sequence. Sn denotes the starting node; blue dashed line indicates observation history; green dashed lines represent ground truth. Interpretations for INT[_] and LC[_] are as per [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗

read the original abstract

We present a general answer set programming based hybrid quantitative-qualitative method for computing constrained branching trajectory modes for moving objects in real-world settings. The method performs constrained traversal of an environment graph, enumerating geometrically admissible motion behaviours as stable models, each constituting a distinct trajectory mode characterised by both domain-dependent and independent factors such as derived event sequence, map topology, and domain norms. The hybrid trajectory computation method is generally applicable across motion characteristics typically encountered in diverse dynamic domains with moving objects, e.g., autonomous driving. We demonstrate applicability and highlight how computed trajectories are traceable to their underlying stable model, thereby affording verifiable interpretability that purely learned approaches cannot provide. We also perform an empirical evaluation with Argoverse 2, a large-scale real-world autonomous driving benchmark representative of the class of dynamic domains within the scope of the proposed method.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

ASP-based trajectory enumeration is a reasonable idea with traceable models, but the generality claim rests on a single driving benchmark and unshown encoding effort.

read the letter

The core contribution is a hybrid ASP method that turns environment graphs and domain rules into stable models, each representing a distinct constrained trajectory mode. This gives explicit traceability from output back to the rules, which is useful for verification in settings where black-box learned trajectories are hard to audit.

What works is the framing: it combines quantitative geometry with qualitative norms and event sequences inside answer set programming, and the Argoverse 2 evaluation shows it can produce admissible paths on real driving data. The interpretability angle is genuine; you can inspect which rules fired for a given mode.

The soft spot is the generality claim. The abstract and stress-test note say the method applies across diverse dynamic domains, yet the only reported results are on autonomous driving. No other domains, no discussion of how much rule rewriting is needed when topology or norms change, and no solver scaling data outside this benchmark. If encoding effort or performance degrades outside planar road networks, the "general foundation" label does not follow from the evidence shown.

The math itself looks standard ASP plus graph traversal; nothing obviously circular or invented. Citations are not detailed here, but the approach sits in the existing ASP-for-robotics line rather than claiming to invent the formalism.

This is for people already working on symbolic planning for motion or hybrid neuro-symbolic systems who want an alternative to pure learning. It is worth sending to peer review because the method is concrete and the evaluation is on a public benchmark, even if the scope needs tightening. A referee can check whether the encoding transfers or whether the single-domain result is the real limit.

Referee Report

2 major / 1 minor

Summary. The paper presents a hybrid answer set programming (ASP) method for computing constrained branching trajectory modes of moving objects. It encodes geometric constraints, map topology, and domain norms as ASP rules whose stable models correspond to admissible trajectory modes characterized by event sequences and other factors. The method is positioned as generally applicable to diverse dynamic domains (e.g., autonomous driving) and is evaluated on the Argoverse 2 benchmark, with emphasis on traceability to stable models for interpretability.

Significance. If the central claims hold, the work would provide a verifiable symbolic foundation for motion planning that integrates qualitative norms with quantitative geometry, offering advantages in safety-critical settings over black-box learned methods. The explicit use of stable models for traceability is a notable strength for explainability.

major comments (2)

[Abstract] Abstract: The claim that the hybrid method 'is generally applicable across motion characteristics typically encountered in diverse dynamic domains' is not supported by the reported evaluation, which is confined to the Argoverse 2 autonomous-driving benchmark. No results or encodings are shown for other domains (e.g., non-planar topology or continuous-time norms), so the transferability of the ASP rules cannot be assessed.
[Evaluation] Evaluation section: The paper states that Argoverse 2 is 'representative of the class of dynamic domains within the scope,' yet provides no cross-domain experiments, encoding-effort metrics, or solver-performance comparisons outside driving. This leaves the load-bearing generality assumption untested.

minor comments (1)

[Abstract] The abstract and introduction would benefit from a clearer statement of the precise scope limitations (e.g., planarity assumptions) to avoid overgeneralization.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed comments on the generality claims. We address each point below and indicate where revisions will be made to better align the manuscript with the reported evaluation.

read point-by-point responses

Referee: [Abstract] Abstract: The claim that the hybrid method 'is generally applicable across motion characteristics typically encountered in diverse dynamic domains' is not supported by the reported evaluation, which is confined to the Argoverse 2 autonomous-driving benchmark. No results or encodings are shown for other domains (e.g., non-planar topology or continuous-time norms), so the transferability of the ASP rules cannot be assessed.

Authors: We agree that the empirical support is limited to Argoverse 2 and that explicit encodings or results for other domains are absent. The manuscript's generality claim rests on the modular structure of the ASP encoding (geometric constraints, topology, and norms as separate rule sets) rather than cross-domain experiments. To avoid overstatement, we will revise the abstract to state that the method is designed for adaptability across domains with the provided Argoverse 2 evaluation serving as an initial demonstration on a representative large-scale benchmark. revision: yes
Referee: [Evaluation] Evaluation section: The paper states that Argoverse 2 is 'representative of the class of dynamic domains within the scope,' yet provides no cross-domain experiments, encoding-effort metrics, or solver-performance comparisons outside driving. This leaves the load-bearing generality assumption untested.

Authors: The evaluation section justifies Argoverse 2 on the basis of its scale and coverage of complex motion patterns, but we accept that this does not constitute a test of transferability. We will add a short discussion subsection clarifying that the ASP rules are intended to be reusable by substituting domain-specific predicates, while noting that quantitative transferability metrics remain future work. This revision will temper the representative claim without altering the core technical contribution. revision: yes

Circularity Check

0 steps flagged

No circularity detected; derivation self-contained

full rationale

The paper presents an ASP-based hybrid method where constraints are encoded as rules and stable models are taken to represent admissible trajectories by the semantics of answer-set programming. No equations, fitted parameters, or self-citations are exhibited that reduce any claimed prediction or generality result to the inputs by construction. The central claim rests on the expressiveness of ASP encodings for domain norms and geometry, which is an independent modeling choice rather than a definitional loop or renamed fit. Evaluation on Argoverse 2 is presented as demonstration, not as the source of the generality assertion.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that ASP can capture both domain-dependent and independent factors through graph traversal and stable models; no free parameters or invented entities are mentioned in the abstract.

axioms (1)

domain assumption The environment can be represented as a graph for constrained traversal where stable models correspond to admissible trajectories
The method performs constrained traversal of an environment graph, enumerating geometrically admissible motion behaviours as stable models.

pith-pipeline@v0.9.1-grok · 5679 in / 1076 out tokens · 22467 ms · 2026-06-25T20:40:15.552392+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

22 extracted references · 10 canonical work pages

[1]

In: ACAI

Bhatt, M., Suchan, J.: Artificial visual intelligence: Perceptual commonsense for human-centred cognitive technologies. In: ACAI. LNCS, vol. 13500, pp. 216–242. Springer (2021). https://doi.org/10.1007/978-3-031-24349-3_12

work page doi:10.1007/978-3-031-24349-3_12 2021
[2]

https://doi.org/10.1145/2043174

Brewka,G.,Eiter,T.,Truszczyński,M.:Answersetprogrammingataglance.Com- munications of the ACM54(12), 92–103 (2011). https://doi.org/10.1145/2043174. 2043195

work page doi:10.1145/2043174 2011
[3]

In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Chang, M.F., Lambert, J., Sangkloy, P., Singh, J., Bak, S., Hartnett, A., Wang, D., Carr, P., Lucey, S., Ramanan, D., Hays, J.: Argoverse: 3d tracking and forecasting with rich maps. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 8740–8749 (2019)

2019
[4]

12 Compositional Behavioral Semantics Moody, J., Wu, L., Liao, Y ., and Saffell, M

Davis, E.: Logical formalizations of commonsense reasoning: A survey. Journal of Artificial Intelligence Research59, 651–723 (2017). https://doi.org/10.1613/jair. 5339

work page doi:10.1613/jair 2017
[5]

Communications of the ACM58(9), 92–103 (2015)

Davis, E., Marcus, G.: Commonsense reasoning and commonsense knowledge in artificial intelligence. Communications of the ACM58(9), 92–103 (2015). https: //doi.org/10.1145/2701413 Computing Reasonable Motion17

work page doi:10.1145/2701413 2015
[6]

European Commission: Proposal for a regulation laying down harmonised rules on artificial intelligence (artificial intelligence act) (2021)

2021
[7]

CoRRabs/1405.3694(2014)

Gebser, M., Kaminski, R., Kaufmann, B., Schaub, T.: Clingo = ASP + control: Preliminary report. CoRRabs/1405.3694(2014)

Pith/arXiv arXiv 2014
[8]

Khandelwal, S., Qi, W., Singh, J., Hartnett, A., Ramanan, D.: What-if motion prediction for autonomous driving (2020), https://arxiv.org/abs/2008.10587

arXiv 2020
[9]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

Knoche, M., de Geus, D., Leibe, B.: DONUT: A decoder-only model for trajec- tory prediction. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 28903–28912 (2025)

2025
[10]

Springer (2019)

Lifschitz, V.: Answer Set Programming. Springer (2019). https://doi.org/10.1007/ 978-3-030-24658-7

2019
[11]

Understanding and

Mangalam, K., An, Y., Girase, H., Malik, J.: From goals, waypoints & paths to long term human trajectory forecasting. In: ICCV. pp. 15213–15222 (2021). https: //doi.org/10.1109/ICCV48922.2021.01495

work page doi:10.1109/iccv48922.2021.01495 2021
[12]

In: Rules and Reasoning: 9th International Joint Conference, RuleML+RR 2025, Istanbul, Turkey, Septem- ber 22–24, 2025, Proceedings

Monsen, J., Suchan, J., Bhatt, M.: Probabilistic answer set programming driven ranking of dynamic space-time belief models. In: Rules and Reasoning: 9th International Joint Conference, RuleML+RR 2025, Istanbul, Turkey, Septem- ber 22–24, 2025, Proceedings. p. 156–175. Springer-Verlag, Berlin, Heidelberg (2025). https://doi.org/10.1007/978-3-032-08887-1_10...

work page doi:10.1007/978-3-032-08887-1_10 2025
[13]

Shi, J., Chen, J., Wang, Y., Sun, L., Liu, C., Xiong, W., Wo, T.: Motion forecasting for autonomous vehicles: A survey (2025), https://arxiv.org/abs/2502.08664

arXiv 2025
[14]

Siméoni, O., Vo, H.V., Seitzer, M., Baldassarre, F., Oquab, M., Jose, C., Khalidov, V., Szafraniec, M., Yi, S., Ramamonjisoa, M., Massa, F., Haziza, D., Wehrstedt, L., Wang, J., Darcet, T., Moutakanni, T., Sentana, L., Roberts, C., Vedaldi, A., Tolan, J., Brandt, J., Couprie, C., Mairal, J., Jégou, H., Labatut, P., Bojanowski, P.: Dinov3 (2025), https://a...

Pith/arXiv arXiv 2025
[15]

In: Foundations of Information and Knowl- edge Systems: 14th International Symposium, FoIKS 2026, Hanover, Germany, March 23–26, 2026, Proceedings

Suchan, J., Baloch, S., Bhatt, M.: Towards a VLM-based foundation for generalised neurosymbolic visual commonsense. In: Foundations of Information and Knowl- edge Systems: 14th International Symposium, FoIKS 2026, Hanover, Germany, March 23–26, 2026, Proceedings. p. 359–365. Springer-Verlag, Berlin, Heidelberg (2026). https://doi.org/10.1007/978-3-032-215...

work page doi:10.1007/978-3-032-21540-6_24 2026
[16]

ASP-driven visual commonsense: a general framework for reasoning about embodied interaction in the wild

Suchan, J., Bhatt, M., Monsen, J.: ASP-driven visual commonsense: a general framework for reasoning about embodied interaction in the wild. In: Proceedings of the 22nd International Conference on Principles of Knowledge Representation and Reasoning. KR ’25 (2025). https://doi.org/10.24963/kr.2025/61, https://doi. org/10.24963/kr.2025/61

work page doi:10.24963/kr.2025/61 2025
[17]

Suchan, J., Bhatt, M., Varadarajan, S.: Commonsense visual sensemaking for au- tonomous driving - on generalised neurosymbolic online abduction integrating vi- sion and semantics. Artif. Intell.299, 103522 (2021). https://doi.org/10.1016/J. ARTINT.2021.103522

work page doi:10.1016/j 2021
[18]

The reflectance field map: Mapping glass and specular surfaces in dynamic environment s,

Wang, M., Zhu, X., Yu, C., Li, W., Ma, Y., Jin, R., Ren, X., Ren, D., Wang, M., Yang, W.: Ganet: Goal area network for motion forecasting. In: ICRA. pp. 1609–1615 (2023). https://doi.org/10.1109/ICRA48891.2023.10160468

work page doi:10.1109/icra48891.2023.10160468 2023
[19]

In: NeurIPS Datasets and Benchmarks (2021) 18 Monsen et al

Wilson, B., Qi, W., Agarwal, T., Lambert, J., Singh, J., Khandelwal, S., Pan, B., Kumar, R., Hartnett, A., Pontes, J.K., Ramanan, D., Carr, P., Hays, J.: Argoverse 2: Next generation datasets for self-driving perception and forecasting. In: NeurIPS Datasets and Benchmarks (2021) 18 Monsen et al

2021
[20]

IEEE Transactions on Intelligent Transportation Systems25(7), 6770–6779 (2024)

Yao, Z., Li, X., Lang, B., Chuah, M.C.: Goal-lbp: Goal-based local behavior guided trajectory prediction for autonomous driving. IEEE Transactions on Intelligent Transportation Systems25(7), 6770–6779 (2024)

2024
[21]

In: The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

Zhang, B., Song, N., Zhang, L.: DeMo: Decoupling motion forecasting into direc- tional intentions and dynamic states. In: The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

2024
[22]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Zhou, Z., Wang, J., Li, Y.H., Huang, Y.K.: Query-centric trajectory prediction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 17863–17873 (2023)

2023

[1] [1]

In: ACAI

Bhatt, M., Suchan, J.: Artificial visual intelligence: Perceptual commonsense for human-centred cognitive technologies. In: ACAI. LNCS, vol. 13500, pp. 216–242. Springer (2021). https://doi.org/10.1007/978-3-031-24349-3_12

work page doi:10.1007/978-3-031-24349-3_12 2021

[2] [2]

https://doi.org/10.1145/2043174

Brewka,G.,Eiter,T.,Truszczyński,M.:Answersetprogrammingataglance.Com- munications of the ACM54(12), 92–103 (2011). https://doi.org/10.1145/2043174. 2043195

work page doi:10.1145/2043174 2011

[3] [3]

In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Chang, M.F., Lambert, J., Sangkloy, P., Singh, J., Bak, S., Hartnett, A., Wang, D., Carr, P., Lucey, S., Ramanan, D., Hays, J.: Argoverse: 3d tracking and forecasting with rich maps. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 8740–8749 (2019)

2019

[4] [4]

12 Compositional Behavioral Semantics Moody, J., Wu, L., Liao, Y ., and Saffell, M

Davis, E.: Logical formalizations of commonsense reasoning: A survey. Journal of Artificial Intelligence Research59, 651–723 (2017). https://doi.org/10.1613/jair. 5339

work page doi:10.1613/jair 2017

[5] [5]

Communications of the ACM58(9), 92–103 (2015)

Davis, E., Marcus, G.: Commonsense reasoning and commonsense knowledge in artificial intelligence. Communications of the ACM58(9), 92–103 (2015). https: //doi.org/10.1145/2701413 Computing Reasonable Motion17

work page doi:10.1145/2701413 2015

[6] [6]

European Commission: Proposal for a regulation laying down harmonised rules on artificial intelligence (artificial intelligence act) (2021)

2021

[7] [7]

CoRRabs/1405.3694(2014)

Gebser, M., Kaminski, R., Kaufmann, B., Schaub, T.: Clingo = ASP + control: Preliminary report. CoRRabs/1405.3694(2014)

Pith/arXiv arXiv 2014

[8] [8]

Khandelwal, S., Qi, W., Singh, J., Hartnett, A., Ramanan, D.: What-if motion prediction for autonomous driving (2020), https://arxiv.org/abs/2008.10587

arXiv 2020

[9] [9]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

Knoche, M., de Geus, D., Leibe, B.: DONUT: A decoder-only model for trajec- tory prediction. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 28903–28912 (2025)

2025

[10] [10]

Springer (2019)

Lifschitz, V.: Answer Set Programming. Springer (2019). https://doi.org/10.1007/ 978-3-030-24658-7

2019

[11] [11]

Understanding and

Mangalam, K., An, Y., Girase, H., Malik, J.: From goals, waypoints & paths to long term human trajectory forecasting. In: ICCV. pp. 15213–15222 (2021). https: //doi.org/10.1109/ICCV48922.2021.01495

work page doi:10.1109/iccv48922.2021.01495 2021

[12] [12]

In: Rules and Reasoning: 9th International Joint Conference, RuleML+RR 2025, Istanbul, Turkey, Septem- ber 22–24, 2025, Proceedings

Monsen, J., Suchan, J., Bhatt, M.: Probabilistic answer set programming driven ranking of dynamic space-time belief models. In: Rules and Reasoning: 9th International Joint Conference, RuleML+RR 2025, Istanbul, Turkey, Septem- ber 22–24, 2025, Proceedings. p. 156–175. Springer-Verlag, Berlin, Heidelberg (2025). https://doi.org/10.1007/978-3-032-08887-1_10...

work page doi:10.1007/978-3-032-08887-1_10 2025

[13] [13]

Shi, J., Chen, J., Wang, Y., Sun, L., Liu, C., Xiong, W., Wo, T.: Motion forecasting for autonomous vehicles: A survey (2025), https://arxiv.org/abs/2502.08664

arXiv 2025

[14] [14]

Siméoni, O., Vo, H.V., Seitzer, M., Baldassarre, F., Oquab, M., Jose, C., Khalidov, V., Szafraniec, M., Yi, S., Ramamonjisoa, M., Massa, F., Haziza, D., Wehrstedt, L., Wang, J., Darcet, T., Moutakanni, T., Sentana, L., Roberts, C., Vedaldi, A., Tolan, J., Brandt, J., Couprie, C., Mairal, J., Jégou, H., Labatut, P., Bojanowski, P.: Dinov3 (2025), https://a...

Pith/arXiv arXiv 2025

[15] [15]

In: Foundations of Information and Knowl- edge Systems: 14th International Symposium, FoIKS 2026, Hanover, Germany, March 23–26, 2026, Proceedings

Suchan, J., Baloch, S., Bhatt, M.: Towards a VLM-based foundation for generalised neurosymbolic visual commonsense. In: Foundations of Information and Knowl- edge Systems: 14th International Symposium, FoIKS 2026, Hanover, Germany, March 23–26, 2026, Proceedings. p. 359–365. Springer-Verlag, Berlin, Heidelberg (2026). https://doi.org/10.1007/978-3-032-215...

work page doi:10.1007/978-3-032-21540-6_24 2026

[16] [16]

ASP-driven visual commonsense: a general framework for reasoning about embodied interaction in the wild

Suchan, J., Bhatt, M., Monsen, J.: ASP-driven visual commonsense: a general framework for reasoning about embodied interaction in the wild. In: Proceedings of the 22nd International Conference on Principles of Knowledge Representation and Reasoning. KR ’25 (2025). https://doi.org/10.24963/kr.2025/61, https://doi. org/10.24963/kr.2025/61

work page doi:10.24963/kr.2025/61 2025

[17] [17]

Suchan, J., Bhatt, M., Varadarajan, S.: Commonsense visual sensemaking for au- tonomous driving - on generalised neurosymbolic online abduction integrating vi- sion and semantics. Artif. Intell.299, 103522 (2021). https://doi.org/10.1016/J. ARTINT.2021.103522

work page doi:10.1016/j 2021

[18] [18]

The reflectance field map: Mapping glass and specular surfaces in dynamic environment s,

Wang, M., Zhu, X., Yu, C., Li, W., Ma, Y., Jin, R., Ren, X., Ren, D., Wang, M., Yang, W.: Ganet: Goal area network for motion forecasting. In: ICRA. pp. 1609–1615 (2023). https://doi.org/10.1109/ICRA48891.2023.10160468

work page doi:10.1109/icra48891.2023.10160468 2023

[19] [19]

In: NeurIPS Datasets and Benchmarks (2021) 18 Monsen et al

Wilson, B., Qi, W., Agarwal, T., Lambert, J., Singh, J., Khandelwal, S., Pan, B., Kumar, R., Hartnett, A., Pontes, J.K., Ramanan, D., Carr, P., Hays, J.: Argoverse 2: Next generation datasets for self-driving perception and forecasting. In: NeurIPS Datasets and Benchmarks (2021) 18 Monsen et al

2021

[20] [20]

IEEE Transactions on Intelligent Transportation Systems25(7), 6770–6779 (2024)

Yao, Z., Li, X., Lang, B., Chuah, M.C.: Goal-lbp: Goal-based local behavior guided trajectory prediction for autonomous driving. IEEE Transactions on Intelligent Transportation Systems25(7), 6770–6779 (2024)

2024

[21] [21]

In: The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

Zhang, B., Song, N., Zhang, L.: DeMo: Decoupling motion forecasting into direc- tional intentions and dynamic states. In: The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS) (2024)

2024

[22] [22]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Zhou, Z., Wang, J., Li, Y.H., Huang, Y.K.: Query-centric trajectory prediction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 17863–17873 (2023)

2023