pith. machine review for the scientific record. sign in

arxiv: 2604.17679 · v1 · submitted 2026-04-20 · 💻 cs.RO

Recognition: unknown

A Hamilton-Jacobi Reachability-Guided Search Framework for Efficient and Safe Indoor Planar Robot Navigation

Authors on Pith no claims yet

Pith reviewed 2026-05-10 05:18 UTC · model grok-4.3

classification 💻 cs.RO
keywords robot navigationHamilton-Jacobi reachabilitygraph searchpath planningsafety constraintsheuristicsindoor environmentsdynamic obstacles
0
0 comments X

The pith

Precomputed Hamilton-Jacobi reachability value functions guide online graph search to make robot navigation both faster and safer.

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

The paper sets out to improve autonomous indoor robot navigation so that paths can be found quickly while avoiding collisions even when the space contains moving people or unknown obstacles. It does so by computing Hamilton-Jacobi reachability value functions once offline and then feeding those values into a standard graph-search planner at runtime. The precomputed values act both as better-than-default heuristics that prune the search tree and as explicit safety constraints that keep candidate paths away from dangerous states. Because the search itself runs online, the method can react to new sensor data without needing a complete static map in advance. The result is a planner that runs faster than pure graph search yet remains safer than methods that ignore reachability.

Core claim

Precomputed HJ value functions, used as informative heuristics and proactive safety constraints, amortize online computation of the graph search process. At the same time, graph search enables reachability-based reasoning to be incorporated into online planning, overcoming the long-standing challenge of HJ reachability requiring full knowledge of the environment.

What carries the argument

Offline-computed Hamilton-Jacobi reachability value functions inserted into online graph search both to steer the search with admissible heuristics and to enforce proactive safety constraints on candidate paths.

If this is right

  • Online planning time decreases because the precomputed values prune large portions of the search space.
  • Safety improves because unsafe states are excluded before the robot commits to a path.
  • The planner works with only partial or incrementally revealed environment information.
  • Performance gains appear in both simulation and physical indoor experiments with and without humans.

Where Pith is reading between the lines

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

  • The same offline-online split could be tried with other reachability or optimal-control precomputations in 3D or non-planar robot models.
  • If the environment changes faster than the precomputation can be refreshed, an incremental update scheme for the value functions would be needed.
  • The approach might transfer to other graph-search domains such as motion planning for manipulators or multi-robot coordination where safety margins are critical.

Load-bearing premise

The precomputed reachability values remain accurate enough to guide search and enforce safety even when the real environment differs from the one used for precomputation or contains moving obstacles.

What would settle it

Deploy the planner in a controlled indoor testbed with moving human obstacles, measure wall-clock planning time and collision rate against standard A* and pure HJ baselines, and check whether the hybrid method loses its reported advantage in either metric.

Figures

Figures reproduced from arXiv: 2604.17679 by Cameron Siu, Hanyang Hu, Mo Chen.

Figure 1
Figure 1. Figure 1: Overview of the proposed HJ reachability-guided [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Heuristic comparisons for the robot at two orientations: [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: , the volume of the BRT increases with the robot’s speed (positive z-axis direction), reflecting the reduced ability of the robot to avoid collisions at higher velocities. This pruning strategy is particularly effective for static obstacles. Since ob￾stacle configurations remain unchanged over time, dangerous nodes can be identified and pruned as soon as their BRT values approach the safety threshold ϵmap,… view at source ↗
Figure 4
Figure 4. Figure 4: The influence of heading angle and relative distance [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Simulated navigation costmap with static obstacles. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Left: ANA* w/ Dist created a trajectory requiring con￾tinuous backward movement (The heading angles represented by arrows are always pointing downwards). Right: ANA* w/ TTR generated a more reasonable trajectory where the robot reversed briefly and reoriented toward the goal. B. Pruning Strategies Comparison BRT pruner improves the search efficiency and trajectory safety by cutting off the nodes that will … view at source ↗
Figure 7
Figure 7. Figure 7: Left: ANA* w/ OBS pruner got too close to the obstacle to find the goal; Right: ANA* w/ BRT pruner eliminated dangerous nodes early and found the goal. As for ANA*-based planners, we compared the effects of two pruning strategies under the same long-distance setting with 12 initial groups. As shown in Table IV, ANA* equipped with the OBS pruner failed in 4 out of the 12 tasks, whereas the version with the … view at source ↗
Figure 8
Figure 8. Figure 8: A sequence from the Face-to-Face scenario with the [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Layout of the two static environments: Long Hallway [PITH_FULL_IMAGE:figures/full_fig_p014_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Comparison of the OBS and BRT pruners across consecutive timesteps (from [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Layout of the dynamic environment. Dotted lines [PITH_FULL_IMAGE:figures/full_fig_p015_11.png] view at source ↗
read the original abstract

Autonomous navigation requires planning to reach a goal safely and efficiently in complex and potentially dynamic environments. Graph search-based algorithms are widely adopted due to their generality and theoretical guarantees when equipped with admissible heuristics. However, the computational complexity of graph search grows rapidly with the dimensionality of the search space, often making real-time planning in dynamic environments intractable. In this paper, we combine offline Hamilton-Jacobi (HJ) reachability with online graph search to leverage the complementary strengths of both. Precomputed HJ value functions, used as informative heuristics and proactive safety constraints, amortize online computation of the graph search process. At the same time, graph search enables reachability-based reasoning to be incorporated into online planning, overcoming the long-standing challenge of HJ reachability requiring full knowledge of the environment. Extensive simulation studies and real-world experiments demonstrate that the proposed approach consistently outperforms baseline methods in terms of planning efficiency and navigation safety, in environments with and without human presence.

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.

Referee Report

2 major / 2 minor

Summary. The paper proposes a hybrid navigation framework for planar indoor robots that precomputes Hamilton-Jacobi (HJ) reachability value functions offline and uses them online as informative heuristics and proactive safety constraints within a graph-search planner. This combination is claimed to amortize online computation while incorporating reachability-based safety reasoning, thereby overcoming HJ reachability's traditional requirement for full environment knowledge. The approach is evaluated in simulations and real-world experiments (with and without humans) and asserted to consistently outperform baseline methods in planning efficiency and navigation safety.

Significance. If the integration is sound and the experimental claims hold, the work would be significant for bridging offline safety analysis with online adaptability in robot motion planning. It directly targets the computational intractability of high-dimensional graph search and the full-knowledge limitation of HJ methods, offering a practical path toward safer real-time navigation in dynamic indoor spaces. The use of precomputed value functions to guide search while retaining graph-search generality is a conceptually clean contribution.

major comments (2)
  1. [Abstract and §1] Abstract and §1 (Introduction): The central claim that graph search 'overcomes the long-standing challenge of HJ reachability requiring full knowledge of the environment' requires clarification. Precomputing the HJ value functions still demands a complete differential-game model (workspace, obstacles, robot dynamics) at offline time. The manuscript must explicitly address how the method remains safe and effective when the online environment deviates from this model (e.g., unmodeled moving humans, partial observability, or layout changes), and whether the safety constraints remain valid or become conservative over-approximations.
  2. [§4] §4 (Experimental Evaluation) and associated tables/figures: The abstract states that the method 'consistently outperforms baseline methods' in efficiency and safety, yet no quantitative metrics, baseline descriptions, success rates, timing distributions, or statistical significance tests are referenced. Without these details it is impossible to evaluate whether the hybrid approach delivers practically meaningful gains or merely marginal improvements under the tested conditions.
minor comments (2)
  1. [§3] Notation for the HJ value function and its embedding into the graph-search heuristic should be defined more explicitly (e.g., how the value function is discretized and queried during A* or similar search).
  2. [§2] The manuscript should include a clear statement of the assumptions on environment dynamics (static vs. moving obstacles) and whether the precomputed HJ functions are recomputed or adapted online.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed and insightful comments on our manuscript. We have carefully considered each point and provide point-by-point responses below. Where appropriate, we have made revisions to clarify our contributions and strengthen the experimental presentation.

read point-by-point responses
  1. Referee: [Abstract and §1] Abstract and §1 (Introduction): The central claim that graph search 'overcomes the long-standing challenge of HJ reachability requiring full knowledge of the environment' requires clarification. Precomputing the HJ value functions still demands a complete differential-game model (workspace, obstacles, robot dynamics) at offline time. The manuscript must explicitly address how the method remains safe and effective when the online environment deviates from this model (e.g., unmodeled moving humans, partial observability, or layout changes), and whether the safety constraints remain valid or become conservative over-approximations.

    Authors: We thank the referee for highlighting this important clarification. The precomputation indeed requires a complete nominal model offline. However, the integration with graph search allows the planner to use online sensor information to update the search graph and replan in real time, thereby handling deviations from the nominal model such as moving humans or layout changes without needing full knowledge at runtime. The safety constraints from the HJ value functions are used as conservative over-approximations in the online phase; they ensure collision avoidance with respect to the modeled obstacles but may result in more conservative paths when unmodeled elements are present. We have revised §1 to include an explicit discussion of these points, the assumptions made, and the conditions under which the safety guarantees hold or become conservative. This addresses the central claim more precisely. revision: yes

  2. Referee: [§4] §4 (Experimental Evaluation) and associated tables/figures: The abstract states that the method 'consistently outperforms baseline methods' in efficiency and safety, yet no quantitative metrics, baseline descriptions, success rates, timing distributions, or statistical significance tests are referenced. Without these details it is impossible to evaluate whether the hybrid approach delivers practically meaningful gains or merely marginal improvements under the tested conditions.

    Authors: We agree that the abstract does not reference the specific quantitative results. In §4, we provide detailed experimental results including baseline descriptions (A* with different heuristics, sampling-based methods), success rates across multiple trials, timing distributions shown in figures, path efficiency metrics, and safety statistics (e.g., minimum distances). Statistical significance is evaluated using paired t-tests. To improve clarity, we have revised the abstract to briefly mention key performance improvements and added references to the relevant tables and figures in the introduction. We have also ensured all quantitative details are prominently presented in §4. These changes allow readers to better assess the practical significance of the gains. revision: yes

Circularity Check

0 steps flagged

No circularity: algorithmic hybrid with independent components

full rationale

The paper describes a practical combination of offline precomputed HJ value functions (as heuristics and safety constraints) with online graph search. No derivation chain, equations, or theorems are presented that reduce by construction to their own inputs. The central claim is an engineering integration that amortizes computation and relaxes full-environment requirements at runtime; this is supported by simulations and experiments rather than self-referential logic. No self-citations, ansatzes, or renamings function as load-bearing steps in any mathematical sense.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The paper is an algorithmic proposal that builds on established HJ reachability and graph-search methods; the abstract introduces no new free parameters, mathematical axioms, or invented entities.

pith-pipeline@v0.9.0 · 5466 in / 1144 out tokens · 54954 ms · 2026-05-10T05:18:01.206739+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

50 extracted references · 7 canonical work pages

  1. [1]

    Rapidly-exploring random trees: A new tool for path planning,

    S. LaValle, “Rapidly-exploring random trees: A new tool for path planning,”Research Report 9811, 1998

  2. [2]

    Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,

    L. E. Kavraki, P. Svestka, J.-C. Latombe, and M. H. Overmars, “Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,”IEEE transactions on Robotics and Automation, vol. 12, no. 4, pp. 566–580, 1996

  3. [3]

    Sampling-based motion planning: A comparative review,

    A. Orthey, C. Chamzas, and L. E. Kavraki, “Sampling-based motion planning: A comparative review,”Annual Review of Control, Robotics, and Autonomous Systems, vol. 7, 2023

  4. [4]

    Sampling-based algorithms for optimal motion planning,

    S. Karaman and E. Frazzoli, “Sampling-based algorithms for optimal motion planning,”The international journal of robotics research, vol. 30, no. 7, pp. 846–894, 2011

  5. [5]

    Fast marching tree: A fast marching sampling-based method for optimal motion planning in many dimensions,

    L. Janson, E. Schmerling, A. Clark, and M. Pavone, “Fast marching tree: A fast marching sampling-based method for optimal motion planning in many dimensions,”The International journal of robotics research, vol. 34, no. 7, pp. 883–921, 2015

  6. [6]

    Review on model predic- tive control: An engineering perspective,

    M. Schwenzer, M. Ay, T. Bergs, and D. Abel, “Review on model predic- tive control: An engineering perspective,”The International Journal of Advanced Manufacturing Technology, vol. 117, no. 5, pp. 1327–1349, 2021

  7. [7]

    Deep reinforcement learning based mobile robot navigation: A review,

    K. Zhu and T. Zhang, “Deep reinforcement learning based mobile robot navigation: A review,”Tsinghua Science and Technology, vol. 26, no. 5, pp. 674–691, 2021

  8. [8]

    Recent progress in the design and analysis of admissible heuristic functions,

    R. E. Korf, “Recent progress in the design and analysis of admissible heuristic functions,” inInternational Symposium on Abstraction, Refor- mulation, and Approximation, pp. 45–55, Springer, 2000

  9. [9]

    A formal basis for the heuristic determination of minimum cost paths,

    P. E. Hart, N. J. Nilsson, and B. Raphael, “A formal basis for the heuristic determination of minimum cost paths,”IEEE transactions on Systems Science and Cybernetics, vol. 4, no. 2, pp. 100–107, 1968

  10. [10]

    Optimal and efficient path planning for partially-known en- vironments,

    A. Stentz, “Optimal and efficient path planning for partially-known en- vironments,” inProceedings of the 1994 IEEE international conference on robotics and automation, pp. 3310–3317, IEEE, 1994

  11. [11]

    Anytime nonparametric a,

    J. Van Den Berg, R. Shah, A. Huang, and K. Goldberg, “Anytime nonparametric a,” inProceedings of the AAAI conference on artificial intelligence, vol. 25, pp. 105–111, 2011

  12. [12]

    Anytime heuristic search,

    E. A. Hansen and R. Zhou, “Anytime heuristic search,”Journal of Artificial Intelligence Research, vol. 28, pp. 267–297, 2007

  13. [13]

    Hamilton-jacobi reachability: A brief overview and recent advances,

    S. Bansal, M. Chen, S. Herbert, and C. J. Tomlin, “Hamilton-jacobi reachability: A brief overview and recent advances,” in2017 IEEE 56th Annual Conference on Decision and Control (CDC), pp. 2242–2253, IEEE, 2017

  14. [14]

    Hamilton–jacobi reachability: Some recent theoretical advances and applications in unmanned airspace manage- ment,

    M. Chen and C. J. Tomlin, “Hamilton–jacobi reachability: Some recent theoretical advances and applications in unmanned airspace manage- ment,”Annual Review of Control, Robotics, and Autonomous Systems, vol. 1, no. 1, pp. 333–358, 2018

  15. [15]

    One-shot computation of reachable sets for differential games,

    I. Yang, S. Becker-Weimann, M. J. Bissell, and C. J. Tomlin, “One-shot computation of reachable sets for differential games,” inProceedings of the 16th international conference on Hybrid systems: computation and control, pp. 183–192, 2013

  16. [16]

    Heuristic search viewed as path finding in a graph,

    I. Pohl, “Heuristic search viewed as path finding in a graph,”Artificial intelligence, vol. 1, no. 3-4, pp. 193–204, 1970

  17. [17]

    Anytime dynamic a*: An anytime, replanning algorithm.,

    M. Likhachev, D. I. Ferguson, G. J. Gordon, A. Stentz, and S. Thrun, “Anytime dynamic a*: An anytime, replanning algorithm.,” inICAPS, vol. 5, pp. 262–271, 2005

  18. [18]

    Discrete search leading continuous exploration for kinodynamic motion planning.,

    E. Plaku, L. E. Kavraki, and M. Y . Vardi, “Discrete search leading continuous exploration for kinodynamic motion planning.,” inRobotics: Science and Systems, pp. 326–333, 2007

  19. [19]

    Kinodynamic motion planning with state lattice motion primitives,

    M. Pivtoraiko and A. Kelly, “Kinodynamic motion planning with state lattice motion primitives,” in2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 2172–2179, IEEE, 2011

  20. [20]

    Open-source, cost- aware kinematically feasible planning for mobile and surface robotics,

    S. Macenski, M. Booker, J. Wallace, and T. Fischer, “Open-source, cost- aware kinematically feasible planning for mobile and surface robotics,” arXiv preprint arXiv:2401.13078, 2024

  21. [21]

    arXiv preprint arXiv:1803.00653 , year=

    N. Savinov, A. Dosovitskiy, and V . Koltun, “Semi-parametric topological memory for navigation,”arXiv preprint arXiv:1803.00653, 2018

  22. [22]

    Learning to search with mctsnets,

    A. Guez, T. Weber, I. Antonoglou, K. Simonyan, O. Vinyals, D. Wierstra, R. Munos, and D. Silver, “Learning to search with mctsnets,” in International conference on machine learning, pp. 1822–1831, PMLR, 2018

  23. [23]

    Neural network heuristic functions for classical planning: Bootstrapping and comparison to other methods,

    P. Ferber, F. Geißer, F. Trevizan, M. Helmert, and J. Hoffmann, “Neural network heuristic functions for classical planning: Bootstrapping and comparison to other methods,” inProceedings of the International Conference on Automated Planning and Scheduling, vol. 32, pp. 583– 587, 2022

  24. [24]

    Path planning using neural a* search,

    R. Yonetani, T. Taniai, M. Barekatain, M. Nishimura, and A. Kanezaki, “Path planning using neural a* search,” inInternational conference on machine learning, pp. 12029–12039, PMLR, 2021

  25. [25]

    C- mcts: Safe planning with monte carlo tree search,

    D. Parthasarathy, G. Kontes, A. Plinge, and C. Mutschler, “C- mcts: Safe planning with monte carlo tree search,”arXiv preprint arXiv:2305.16209, 2023

  26. [26]

    Constrained hierarchical monte carlo belief-state planning,

    A. Jamgochian, H. Buurmeijer, K. H. Wray, A. Corso, and M. J. Kochen- derfer, “Constrained hierarchical monte carlo belief-state planning,” in2024 IEEE International Conference on Robotics and Automation (ICRA), pp. 2368–2374, IEEE, 2024

  27. [27]

    Williams

    V . Parimi and B. C. Williams, “Risk-bounded multi-agent visual nav- igation via iterative risk allocation,”arXiv preprint arXiv:2509.08157, 2025

  28. [28]

    A risk-aware path planning strategy for uavs in urban environments,

    S. Primatesta, G. Guglieri, and A. Rizzo, “A risk-aware path planning strategy for uavs in urban environments,”Journal of Intelligent & Robotic Systems, vol. 95, no. 2, pp. 629–643, 2019

  29. [29]

    A safe heuristic path-planning method based on a search strategy,

    X. Yan, X. Zhou, and Q. Luo, “A safe heuristic path-planning method based on a search strategy,”Sensors, vol. 24, no. 1, p. 101, 2023. UNDER REVIEW 17

  30. [30]

    On safety and liveness filtering using hamilton-jacobi reachability analysis,

    J. Borquez, K. Chakraborty, H. Wang, and S. Bansal, “On safety and liveness filtering using hamilton-jacobi reachability analysis,”IEEE Transactions on Robotics, 2024

  31. [31]

    Fastrack: a modular framework for real-time motion planning and guaranteed safe tracking,

    M. Chen, S. L. Herbert, H. Hu, Y . Pu, J. F. Fisac, S. Bansal, S. Han, and C. J. Tomlin, “Fastrack: a modular framework for real-time motion planning and guaranteed safe tracking,”IEEE Transactions on Automatic Control, vol. 66, no. 12, pp. 5861–5876, 2021

  32. [32]

    The safety filter: A unified view of safety-critical control in autonomous systems,

    K.-C. Hsu, H. Hu, and J. F. Fisac, “The safety filter: A unified view of safety-critical control in autonomous systems,”Annual Review of Control, Robotics, and Autonomous Systems, vol. 7, 2023

  33. [33]

    Safe learning in the real world via adaptive shielding with hamilton-jacobi reachability,

    M. Lu, J. S. Gosain, L. Sang, and M. Chen, “Safe learning in the real world via adaptive shielding with hamilton-jacobi reachability,” Proceedings of Machine Learning Research vol, vol. 283, pp. 1–14, 2025

  34. [34]

    Uncertainty-aware latent safety filters for avoiding out-of-distribution failures,

    J. Seo, K. Nakamura, and A. Bajcsy, “Uncertainty-aware latent safety filters for avoiding out-of-distribution failures,”arXiv preprint arXiv:2505.00779, 2025

  35. [35]

    2502.00935 , archiveprefix =

    K. Nakamura, L. Peters, and A. Bajcsy, “Generalizing safety beyond collision-avoidance via latent-space reachability analysis,”arXiv preprint arXiv:2502.00935, 2025

  36. [36]

    A time-dependent hamilton-jacobi formulation of reachable sets for continuous dynamic games,

    I. M. Mitchell, A. M. Bayen, and C. J. Tomlin, “A time-dependent hamilton-jacobi formulation of reachable sets for continuous dynamic games,”IEEE Transactions on automatic control, vol. 50, no. 7, pp. 947– 957, 2005

  37. [37]

    Reach-avoid problems with time-varying dynamics, targets and constraints,

    J. F. Fisac, M. Chen, C. J. Tomlin, and S. S. Sastry, “Reach-avoid problems with time-varying dynamics, targets and constraints,” in Proceedings of the 18th international conference on hybrid systems: computation and control, pp. 11–20, 2015

  38. [38]

    Robust sequential trajectory planning under disturbances and adversarial intruder,

    M. Chen, S. Bansal, J. F. Fisac, and C. J. Tomlin, “Robust sequential trajectory planning under disturbances and adversarial intruder,”IEEE Transactions on Control Systems Technology, vol. 27, no. 4, pp. 1566– 1582, 2018

  39. [39]

    Human navigational intent inference with probabilistic and optimal approaches,

    P. Agand, M. Taherahmadi, A. Lim, and M. Chen, “Human navigational intent inference with probabilistic and optimal approaches,” in2022 In- ternational Conference on Robotics and Automation (ICRA), pp. 8562– 8568, IEEE, 2022

  40. [40]

    Ttr-based reward for reinforcement learning with implicit model priors. in 2020 ieee,

    X. Lyu and M. Chen, “Ttr-based reward for reinforcement learning with implicit model priors. in 2020 ieee,” inRSJ International Conference on Intelligent Robots and Systems (IROS), pp. 5484–5489, 2020

  41. [41]

    Time-optimal paths for simple cars with moving obstacles in the hamilton-jacobi formulation,

    C. Parkinson and M. Ceccia, “Time-optimal paths for simple cars with moving obstacles in the hamilton-jacobi formulation,” in2022 American Control Conference (ACC), pp. 2944–2949, IEEE, 2022

  42. [42]

    Edelkamp and S

    S. Edelkamp and S. Schr ¨odl,Heuristic search: theory and applications. Elsevier, 2011

  43. [43]

    D* lite,

    S. Koenig and M. Likhachev, “D* lite,” inEighteenth national confer- ence on Artificial intelligence, pp. 476–483, 2002

  44. [44]

    Viscosity solutions of hamilton- jacobi equations,

    M. G. Crandall and P.-L. Lions, “Viscosity solutions of hamilton- jacobi equations,”Transactions of the American Mathematical Society, vol. 277, no. 1, pp. 1–42, 1983

  45. [45]

    Optimizeddp: An efficient, user-friendly library for optimal control and dynamic programming,

    M. Bui, H. Hu, C. He, M. Lu, G. Giovanis, A. Shriraman, and M. Chen, “Optimizeddp: An efficient, user-friendly library for optimal control and dynamic programming,” 2022

  46. [46]

    Probabilistically safe robot planning with confidence-based human predictions,

    J. F. Fisac, A. Bajcsy, S. L. Herbert, D. Fridovich-Keil, S. Wang, C. J. Tomlin, and A. D. Dragan, “Probabilistically safe robot planning with confidence-based human predictions,”arXiv preprint arXiv:1806.00109, 2018

  47. [47]

    Control barrier functions: Theory and applications,

    A. D. Ames, S. Coogan, M. Egerstedt, G. Notomista, K. Sreenath, and P. Tabuada, “Control barrier functions: Theory and applications,” in2019 18th European control conference (ECC), pp. 3420–3431, Ieee, 2019

  48. [48]

    The marathon 2: A navigation system,

    S. Macenski, F. Mart ´ın, R. White, and J. G. Clavero, “The marathon 2: A navigation system,” in2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 2718–2725, IEEE, 2020

  49. [49]

    Real-time loop closure in 2d lidar slam,

    W. Hess, D. Kohler, H. Rapp, and D. Andor, “Real-time loop closure in 2d lidar slam,” in2016 IEEE international conference on robotics and automation (ICRA), pp. 1271–1278, IEEE, 2016

  50. [50]

    Quanser Inc., Markham, Ontario, Canada, 2023

    Quanser Inc.,QBot Platform User Manual. Quanser Inc., Markham, Ontario, Canada, 2023. Available at https://docs.quanser.com/quarc/ documentation/qbot platform.html