TEACar: An Open-Source Autonomous Driving Platform

Andrew Kappes; Dana Moreno; Ivan Ruchkin; Jack Penn; Maxwell Ruyle; Tyler Ruble; William Shaoul; Zhongzheng Zhang

arxiv: 2604.24934 · v2 · submitted 2026-04-27 · 💻 cs.RO · cs.SY· eess.SY

TEACar: An Open-Source Autonomous Driving Platform

Zhongzheng Zhang , Maxwell Ruyle , Andrew Kappes , Tyler Ruble , William Shaoul , Dana Moreno , Jack Penn , Ivan Ruchkin This is my paper

Pith reviewed 2026-05-12 00:46 UTC · model grok-4.3

classification 💻 cs.RO cs.SYeess.SY

keywords small-scale autonomous vehiclemodular hardware platformROS 2CNN-based controlhardware-in-the-loop testingintelligent transportation systemsopen-source roboticsvision-based perception

0 comments

The pith

TEACAR uses a four-layer deck to create a reconfigurable small-scale car for testing self-driving systems.

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 prove that a 1/14- to 1/16-scale vehicle built with physically separated layers for sensing, computing, actuation, and power can serve as a practical, low-cost platform for validating vision-based and learning-based autonomous driving methods. A sympathetic reader would care because full-size vehicle testing is slow and expensive, so a modular hardware testbed could let more researchers and students run real-world experiments with cameras and neural controllers. The authors built a working prototype, measured its mechanical stability and software timing under three different CNN steering policies, and concluded that the design supports scalable research and education. If the claim holds, labs could iterate on perception and control algorithms faster without large budgets or complex integration work.

Core claim

The central claim is that the four-layer deck architecture physically decouples the subsystems, which improves structural rigidity and makes hardware changes simpler, while the ROS 2 software stack and measured inference latencies demonstrate that the platform can run real-time CNN controllers reliably enough to act as an effective testbed for intelligent transportation systems research.

What carries the argument

The four-layer deck structure that physically separates sensing, computation, actuation, and power subsystems.

If this is right

Labs can swap sensors or computers by moving only one layer instead of rebuilding the chassis.
Educational courses can deploy the same hardware for both basic robotics labs and advanced neural control experiments.
Power and computation budgets stay low enough that multiple vehicles can run simultaneously for multi-agent studies.
The open repository lets other groups reproduce the exact latency and power numbers on their own copies.

Where Pith is reading between the lines

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

The same layered separation idea could be adapted to other small robots that need frequent sensor or actuator changes.
Wider availability of such platforms might shorten the time between new vision algorithms and their hardware validation.
Community contributions through the GitHub repo could add standardized benchmarks that compare different controllers across identical hardware.

Load-bearing premise

That separating the vehicle into four stacked layers actually increases rigidity and reduces reconfiguration time, and that running three CNN steering controllers is enough to show the whole system works for general autonomous driving tasks.

What would settle it

A side-by-side build comparison showing that a conventional single-deck layout takes less time to reconfigure or produces lower vibration levels than the four-layer version, or measurements where the CNN controllers lose stability outside the three tested track conditions.

Figures

Figures reproduced from arXiv: 2604.24934 by Andrew Kappes, Dana Moreno, Ivan Ruchkin, Jack Penn, Maxwell Ruyle, Tyler Ruble, William Shaoul, Zhongzheng Zhang.

**Figure 1.** Figure 1: Overview of the TEACAR platform. driving research, education, and development. The remainder of this paper is organized as follows. Section II reviews related small-scale platforms for autonomous racing. Section III provides an overview of the system and its design goals. Sections IV, V, and VI describe the mechanical, hardware, and software design of TEACAR, respectively. Section VII presents the experim… view at source ↗

**Figure 2.** Figure 2: Overview of the TEACAR deck. the chassis centerline. By isolating variations in wheelbase, mounting hole patterns, and deck elevation to this layer, the remaining deck geometry remains invariant across compatible 1/14- and 1/16-scale chassis. Lower Structural Base. The lower structural base interfaces directly with the adapter layer and supports primary structural loads. This layer houses the main battery … view at source ↗

**Figure 3.** Figure 3: Hardware system architecture. A. Computing Unit The computing unit, the brain of the car, is where software programs run. Its performance is vital for the whole system. Our platform supports multiple NVIDIA® edge computing modules, including the Jetson Xavier NX series, Jetson Orin NX series, and Jetson Orin Nano™ series. These devices offer AI performance ranging from 21 TOPS to 100 TOPS, enabling the exe… view at source ↗

**Figure 4.** Figure 4: Power board architecture view at source ↗

**Figure 5.** Figure 5: Software system architecture A. Controller The controller is responsible for translating user inputs into control messages. This subsystem comprises three nodes: a joystick controller node for manual control and an nn controller node that employs a trained neural network for autonomous driving. A base class Controller is provided for nodes within this subsystem. Child nodes invoke functions implemented by… view at source ↗

**Figure 6.** Figure 6: The track for TEACAR experimentation. Racing Track. We built a track shown in view at source ↗

read the original abstract

Intelligent Transportation Systems (ITS) increasingly rely on vision-based perception and learning-based control, necessitating experimental platforms that support realistic hardware-in-the-loop validation. Small-scale platforms for autonomous racing offer a practical path to hardware validation, but often suffer from limited modularity, high integration complexity, or restricted extensibility. This paper presents TEACAR, a 1/14- to 1/16-scale autonomous driving platform designed with modular mechanical architecture, hardware abstraction, and ROS 2-based software. The system adopts a four-layer deck structure that physically decouples sensing, computation, actuation, and power subsystems, improving structural rigidity while simplifying reconfiguration. We constructed and comprehensively evaluated the prototype of TEACAR. Its mechanical stability, structural characteristics, and software performance were quantified based on three CNN-based steering controllers. Inference latency, power consumption, and system operating time were measured to evaluate computational capability and robustness. Our experiments demonstrated that TEACAR offers a scalable, modular, and cost-effective testbed for ITS research, education, and development. Our project repository is available on GitHub.

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

TEACAR adds a four-layer mechanical decoupling for small-scale cars plus ROS 2 abstraction, but three CNN steering runs do not back the scalability and robustness claims.

read the letter

TEACAR is a new 1/14-1/16 scale platform that physically separates sensing, computation, actuation, and power on a four-layer deck, paired with a ROS 2 hardware abstraction layer. That specific mechanical choice and the open GitHub release are the concrete contributions here. The build description is clear enough that someone could replicate the chassis and wiring without too much guesswork, which is useful for labs that want a cheap physical testbed instead of simulation only. The measurements on inference latency, power draw, and run time for three different CNN steering controllers give a basic picture of what the onboard compute can handle under simple control loads. That part is straightforward engineering reporting and worth having in the literature for people who need a starting point. The soft spots sit in the gap between the data and the conclusions. The abstract and stress-test note both flag that the experiments stay inside steering-only CNN inference; there are no results on full perception-planning loops, multi-agent interaction, sensor fusion, or environmental changes like lighting or surface variation. No cost numbers, build times, or head-to-head swap tests versus other small platforms appear, so the adjectives “scalable,” “modular,” and “cost-effective” remain assertions rather than measured outcomes. The four-layer rigidity claim also lacks before-and-after stiffness data or reconfiguration timing. This paper is mainly for educators and small research groups that need an inexpensive, buildable vehicle to run student projects or quick hardware validation. It is not aimed at readers looking for new algorithms or large-scale ITS insights. I would bring it to a reading group only if the group is actively looking for open hardware platforms to try. I would not cite it in my own work unless I were actually adopting the platform. It deserves peer review in a systems venue because the hardware design is reproducible and the code is public, but the authors need to add quantitative comparisons and broader tests before the utility claims hold up.

Referee Report

3 major / 1 minor

Summary. The paper presents TEACAR, a 1/14- to 1/16-scale autonomous driving platform with a modular four-layer deck structure that decouples sensing, computation, actuation, and power subsystems, combined with ROS 2-based software and hardware abstraction. A prototype is constructed and evaluated using three CNN-based steering controllers, with measurements of inference latency, power consumption, and operating time used to quantify mechanical stability, structural characteristics, and software performance. The authors conclude that the platform is scalable, modular, and cost-effective for ITS research, education, and development, with the project released as open-source on GitHub.

Significance. If the modularity and performance claims can be supported with additional quantitative evidence, TEACAR could serve as a practical, low-cost open-source testbed for hardware-in-the-loop validation in vision-based autonomous driving and ITS. The open GitHub repository and ROS 2 integration are explicit strengths that support reproducibility and community use.

major comments (3)

[Abstract] Abstract: the claim that experiments 'demonstrated that TEACAR offers a scalable, modular, and cost-effective testbed' is not entailed by the reported data, which consist only of inference latency, power consumption, and operating time for three steering-only CNN controllers; no cost figures, build times, reconfiguration metrics, or comparisons to existing platforms are provided.
[Evaluation section] Evaluation section (CNN controller tests): restricting the assessment to steering control and computational metrics on three controllers supplies no evidence on perception, planning, multi-agent interaction, or environmental robustness, so the results cannot support the broader claims of scalability and robustness for ITS workloads.
[Mechanical architecture] Mechanical architecture description: the four-layer deck is asserted to improve structural rigidity while simplifying reconfiguration, yet no quantitative data (stiffness measurements, swap-time experiments, or comparisons against alternative designs) are supplied to substantiate these benefits.

minor comments (1)

[Abstract] Abstract: numerical results, error bars, or baseline comparisons for the quantified mechanical stability and software performance are absent, reducing the ability to assess the strength of the experimental claims.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which help improve the clarity and accuracy of our claims. We address each major point below and will revise the manuscript to align the stated conclusions more closely with the presented evidence.

read point-by-point responses

Referee: [Abstract] Abstract: the claim that experiments 'demonstrated that TEACAR offers a scalable, modular, and cost-effective testbed' is not entailed by the reported data, which consist only of inference latency, power consumption, and operating time for three steering-only CNN controllers; no cost figures, build times, reconfiguration metrics, or comparisons to existing platforms are provided.

Authors: We agree that the abstract claim exceeds the scope of the reported experiments. The evaluation quantifies only computational metrics for steering control. We will revise the abstract to state that the experiments demonstrate computational performance and basic platform functionality for steering tasks, while describing the modular architecture as a design feature without claiming quantitative proof of scalability or cost-effectiveness. revision: yes
Referee: [Evaluation section] Evaluation section (CNN controller tests): restricting the assessment to steering control and computational metrics on three controllers supplies no evidence on perception, planning, multi-agent interaction, or environmental robustness, so the results cannot support the broader claims of scalability and robustness for ITS workloads.

Authors: The prototype evaluation is intentionally scoped to steering-only CNN controllers to measure baseline latency, power, and runtime. The ROS 2 architecture and hardware decoupling are intended to enable extension to perception and planning, but the current results do not demonstrate these. We will add a limitations and future-work subsection that explicitly notes the narrow scope of the experiments and outlines how the platform supports broader ITS workloads. revision: partial
Referee: [Mechanical architecture] Mechanical architecture description: the four-layer deck is asserted to improve structural rigidity while simplifying reconfiguration, yet no quantitative data (stiffness measurements, swap-time experiments, or comparisons against alternative designs) are supplied to substantiate these benefits.

Authors: The four-layer deck is presented as a design choice for physical decoupling. No stiffness or timed-reconfiguration measurements were performed. We will revise the mechanical-architecture section to remove unsubstantiated assertions about improved rigidity and simplification, replacing them with a factual description of the layering and its intended benefits, while noting that quantitative validation remains future work. revision: yes

Circularity Check

0 steps flagged

No circularity: descriptive system paper with no derivations or fitted predictions

full rationale

The paper presents a hardware platform description, mechanical design choices, and direct experimental measurements (inference latency, power, operating time) from three CNN steering controllers. No equations, first-principles derivations, parameter fitting, or predictions appear in the provided text or abstract. Claims of scalability/modularity/cost-effectiveness are stated as conclusions from the prototype evaluation rather than derived quantities. No self-citations, ansatzes, or uniqueness theorems are invoked as load-bearing steps. This matches the default expectation for non-circular descriptive work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard robotics engineering practices and the assumption that the described modular design delivers the stated benefits; no free parameters, invented entities, or non-standard axioms are introduced.

axioms (1)

domain assumption ROS 2 middleware is suitable for real-time control and perception in small-scale autonomous vehicles
The software stack is based on ROS 2 without additional justification or performance guarantees provided in the abstract.

pith-pipeline@v0.9.0 · 5510 in / 1239 out tokens · 79664 ms · 2026-05-12T00:46:15.691008+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The system adopts a four-layer deck structure that physically decouples sensing, computation, actuation, and power subsystems... three CNN-based steering controllers... inference latency, power consumption, and system operating time
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Our experiments demonstrated that TEACAR offers a scalable, modular, and cost-effective testbed

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

Intelligent transportation systems: A critical review of integration of cyber-physical systems (cps) and industry 4.0,

M. M. Aslam, W. Shafik, A. F. Hidayatullah, K. Kalinaki, H. Gul, R. Y . Zakari, and A. Tufail, “Intelligent transportation systems: A critical review of integration of cyber-physical systems (cps) and industry 4.0,”Digital Communications and Networks, vol. 12, no. 1, pp. 143–164, 2026. [Online]. Available: https: //www.sciencedirect.com/science/article/pi...

work page 2026
[2]

A review of vision-based multi-task perception research methods for autonomous vehicles,

H. Wang, J. Li, and H. Dong, “A review of vision-based multi-task perception research methods for autonomous vehicles,”Sensors, vol. 25, no. 8, 2025. [Online]. Available: https://www.mdpi.com/1424-8220/25/ 8/2611

work page 2025
[3]

Vision-based environmental perception for autonomous driving,

F. Liu, Z. Lu, and X. Lin, “Vision-based environmental perception for autonomous driving,”Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, vol. 239, no. 1, pp. 39–69, 2025. [Online]. Available: https://doi.org/10.1177/ 09544070231203059

work page 2025
[4]

Research on the application of computer vision based on deep learning in autonomous driving technology,

J. Zhang, J. Cao, J. Chang, X. Li, H. Liu, and Z. Li, “Research on the application of computer vision based on deep learning in autonomous driving technology,” 2024. [Online]. Available: https: //arxiv.org/abs/2406.00490

work page arXiv 2024
[5]

How simulation helps autonomous driving:a survey of sim2real, digital twins, and parallel intelligence,

X. Hu, S. Li, T. Huang, B. Tang, R. Huai, and L. Chen, “How simulation helps autonomous driving:a survey of sim2real, digital twins, and parallel intelligence,” 2023. [Online]. Available: https://arxiv.org/abs/2305.01263

work page arXiv 2023
[6]

The reality gap in robotics: Challenges, solutions, and best practices,

E. Aljalbout, J. Xing, A. Romero, I. Akinola, C. R. Garrett, E. Heiden, A. Gupta, T. Hermans, Y . Narang, D. Fox, D. Scaramuzza, and F. Ramos, “The reality gap in robotics: Challenges, solutions, and best practices,” 2025. [Online]. Available: https://arxiv.org/abs/2510.20808

work page arXiv 2025
[7]

F1/10: An open-source autonomous cyber-physical platform,

M. O’Kelly, V . Sukhil, H. Abbas, J. Harkins, C. Kao, Y . V . Pant, R. Mangharam, D. Agarwal, M. Behl, P. Burgio, and M. Bertogna, “F1/10: An open-source autonomous cyber-physical platform,” 2019. [Online]. Available: https://arxiv.org/abs/1901.08567

work page arXiv 2019
[8]

Teaching autonomous systems hands-on: Leveraging modular small- scale hardware in the robotics classroom,

J. Betz, H. Zheng, Z. Zang, F. Sauerbeck, K. Walas, V . Dimitrov, M. Behl, R. Zheng, J. Biswas, V . Krovi, and R. Mangharam, “Teaching autonomous systems hands-on: Leveraging modular small- scale hardware in the robotics classroom,” 2022. [Online]. Available: https://arxiv.org/abs/2209.11181

work page arXiv 2022
[9]

Donkey car documentation,

“Donkey car documentation,” https://docs.donkeycar.com/, DonkeyCar Project, n.d., accessed: 2026-02-21

work page 2026
[10]

Science Robotics7(66), eabm6074 (2022)

S. Macenski, T. Foote, B. Gerkey, C. Lalancette, and W. Woodall, “Robot operating system 2: Design, architecture, and uses in the wild,” Science Robotics, vol. 7, no. 66, p. eabm6074, 2022. [Online]. Available: https://www.science.org/doi/abs/10.1126/scirobotics.abm6074

work page doi:10.1126/scirobotics.abm6074 2022
[11]

F1tenth: An open-source evaluation environment for continuous control and reinforcement learning,

M. O’Kelly, H. Zheng, D. Karthik, and R. Mangharam, “F1tenth: An open-source evaluation environment for continuous control and reinforcement learning,” inProceedings of the NeurIPS 2019 Competition and Demonstration Track, ser. Proceedings of Machine Learning Research, H. J. Escalante and R. Hadsell, Eds., vol

work page 2019
[12]

PMLR, 08–14 Dec 2020, pp. 77–89. [Online]. Available: https://proceedings.mlr.press/v123/o-kelly20a.html

work page 2020
[13]

Unifying f1tenth autonomous racing: Survey, methods and benchmarks,

B. D. Evans, R. Trumpp, M. Caccamo, F. Jahncke, J. Betz, H. W. Jordaan, and H. A. Engelbrecht, “Unifying f1tenth autonomous racing: Survey, methods and benchmarks,” 2024. [Online]. Available: https://arxiv.org/abs/2402.18558

work page arXiv 2024
[14]

Overcoming the sim- to-real gap in autonomous robots,

P. Trentsios, M. Wolf, and D. Gerhard, “Overcoming the sim- to-real gap in autonomous robots,”Procedia CIRP, vol. 109, pp. 287–292, 2022, 32nd CIRP Design Conference (CIRP Design

work page 2022
[15]

[Online]

- Design in a changing world. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S2212827122007004

work page
[16]

Quantifying the sim2real gap for gps and imu sensors,

I. Mahajan, H. Unjhawala, H. Zhang, Z. Zhou, A. Young, A. Ruiz, S. Caldararu, N. Batagoda, S. Ashokkumar, and D. Negrut, “Quantifying the sim2real gap for gps and imu sensors,” 2024. [Online]. Available: https://arxiv.org/abs/2403.11000

work page arXiv 2024
[17]

Donkeycar: a python self driving library,

DonkeyCar Contributors, “Donkeycar: a python self driving library,” https://github.com/autorope/donkeycar.git, 2026, gitHub repository, ac- cessed Feb. 26, 2026

work page 2026
[18]

Comparison of comprehensive properties of ni-mh (nickel-metal hydride) and li-ion (lithium-ion) batteries in terms of energy efficiency,

J. Kang, F. Yan, P. Zhang, and C. Du, “Comparison of comprehensive properties of ni-mh (nickel-metal hydride) and li-ion (lithium-ion) batteries in terms of energy efficiency,”Energy, vol. 70, pp. 618– 625, 2014. [Online]. Available: https://www.sciencedirect.com/science/ article/pii/S0360544214004514

work page 2014

[1] [1]

Intelligent transportation systems: A critical review of integration of cyber-physical systems (cps) and industry 4.0,

M. M. Aslam, W. Shafik, A. F. Hidayatullah, K. Kalinaki, H. Gul, R. Y . Zakari, and A. Tufail, “Intelligent transportation systems: A critical review of integration of cyber-physical systems (cps) and industry 4.0,”Digital Communications and Networks, vol. 12, no. 1, pp. 143–164, 2026. [Online]. Available: https: //www.sciencedirect.com/science/article/pi...

work page 2026

[2] [2]

A review of vision-based multi-task perception research methods for autonomous vehicles,

H. Wang, J. Li, and H. Dong, “A review of vision-based multi-task perception research methods for autonomous vehicles,”Sensors, vol. 25, no. 8, 2025. [Online]. Available: https://www.mdpi.com/1424-8220/25/ 8/2611

work page 2025

[3] [3]

Vision-based environmental perception for autonomous driving,

F. Liu, Z. Lu, and X. Lin, “Vision-based environmental perception for autonomous driving,”Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, vol. 239, no. 1, pp. 39–69, 2025. [Online]. Available: https://doi.org/10.1177/ 09544070231203059

work page 2025

[4] [4]

Research on the application of computer vision based on deep learning in autonomous driving technology,

J. Zhang, J. Cao, J. Chang, X. Li, H. Liu, and Z. Li, “Research on the application of computer vision based on deep learning in autonomous driving technology,” 2024. [Online]. Available: https: //arxiv.org/abs/2406.00490

work page arXiv 2024

[5] [5]

How simulation helps autonomous driving:a survey of sim2real, digital twins, and parallel intelligence,

X. Hu, S. Li, T. Huang, B. Tang, R. Huai, and L. Chen, “How simulation helps autonomous driving:a survey of sim2real, digital twins, and parallel intelligence,” 2023. [Online]. Available: https://arxiv.org/abs/2305.01263

work page arXiv 2023

[6] [6]

The reality gap in robotics: Challenges, solutions, and best practices,

E. Aljalbout, J. Xing, A. Romero, I. Akinola, C. R. Garrett, E. Heiden, A. Gupta, T. Hermans, Y . Narang, D. Fox, D. Scaramuzza, and F. Ramos, “The reality gap in robotics: Challenges, solutions, and best practices,” 2025. [Online]. Available: https://arxiv.org/abs/2510.20808

work page arXiv 2025

[7] [7]

F1/10: An open-source autonomous cyber-physical platform,

M. O’Kelly, V . Sukhil, H. Abbas, J. Harkins, C. Kao, Y . V . Pant, R. Mangharam, D. Agarwal, M. Behl, P. Burgio, and M. Bertogna, “F1/10: An open-source autonomous cyber-physical platform,” 2019. [Online]. Available: https://arxiv.org/abs/1901.08567

work page arXiv 2019

[8] [8]

Teaching autonomous systems hands-on: Leveraging modular small- scale hardware in the robotics classroom,

J. Betz, H. Zheng, Z. Zang, F. Sauerbeck, K. Walas, V . Dimitrov, M. Behl, R. Zheng, J. Biswas, V . Krovi, and R. Mangharam, “Teaching autonomous systems hands-on: Leveraging modular small- scale hardware in the robotics classroom,” 2022. [Online]. Available: https://arxiv.org/abs/2209.11181

work page arXiv 2022

[9] [9]

Donkey car documentation,

“Donkey car documentation,” https://docs.donkeycar.com/, DonkeyCar Project, n.d., accessed: 2026-02-21

work page 2026

[10] [10]

Science Robotics7(66), eabm6074 (2022)

S. Macenski, T. Foote, B. Gerkey, C. Lalancette, and W. Woodall, “Robot operating system 2: Design, architecture, and uses in the wild,” Science Robotics, vol. 7, no. 66, p. eabm6074, 2022. [Online]. Available: https://www.science.org/doi/abs/10.1126/scirobotics.abm6074

work page doi:10.1126/scirobotics.abm6074 2022

[11] [11]

F1tenth: An open-source evaluation environment for continuous control and reinforcement learning,

M. O’Kelly, H. Zheng, D. Karthik, and R. Mangharam, “F1tenth: An open-source evaluation environment for continuous control and reinforcement learning,” inProceedings of the NeurIPS 2019 Competition and Demonstration Track, ser. Proceedings of Machine Learning Research, H. J. Escalante and R. Hadsell, Eds., vol

work page 2019

[12] [12]

PMLR, 08–14 Dec 2020, pp. 77–89. [Online]. Available: https://proceedings.mlr.press/v123/o-kelly20a.html

work page 2020

[13] [13]

Unifying f1tenth autonomous racing: Survey, methods and benchmarks,

B. D. Evans, R. Trumpp, M. Caccamo, F. Jahncke, J. Betz, H. W. Jordaan, and H. A. Engelbrecht, “Unifying f1tenth autonomous racing: Survey, methods and benchmarks,” 2024. [Online]. Available: https://arxiv.org/abs/2402.18558

work page arXiv 2024

[14] [14]

Overcoming the sim- to-real gap in autonomous robots,

P. Trentsios, M. Wolf, and D. Gerhard, “Overcoming the sim- to-real gap in autonomous robots,”Procedia CIRP, vol. 109, pp. 287–292, 2022, 32nd CIRP Design Conference (CIRP Design

work page 2022

[15] [15]

[Online]

- Design in a changing world. [Online]. Available: https: //www.sciencedirect.com/science/article/pii/S2212827122007004

work page

[16] [16]

Quantifying the sim2real gap for gps and imu sensors,

I. Mahajan, H. Unjhawala, H. Zhang, Z. Zhou, A. Young, A. Ruiz, S. Caldararu, N. Batagoda, S. Ashokkumar, and D. Negrut, “Quantifying the sim2real gap for gps and imu sensors,” 2024. [Online]. Available: https://arxiv.org/abs/2403.11000

work page arXiv 2024

[17] [17]

Donkeycar: a python self driving library,

DonkeyCar Contributors, “Donkeycar: a python self driving library,” https://github.com/autorope/donkeycar.git, 2026, gitHub repository, ac- cessed Feb. 26, 2026

work page 2026

[18] [18]

Comparison of comprehensive properties of ni-mh (nickel-metal hydride) and li-ion (lithium-ion) batteries in terms of energy efficiency,

J. Kang, F. Yan, P. Zhang, and C. Du, “Comparison of comprehensive properties of ni-mh (nickel-metal hydride) and li-ion (lithium-ion) batteries in terms of energy efficiency,”Energy, vol. 70, pp. 618– 625, 2014. [Online]. Available: https://www.sciencedirect.com/science/ article/pii/S0360544214004514

work page 2014