SDVDiag: A Modular Platform for the Diagnosis of Connected Vehicle Functions

Falk Dettinger; Matthias Wei{\ss}; Michael Weyrich

REVIEW 4 major objections 4 minor 24 references

SDVDiag: A Modular Platform for the Diagnosis of Connected Vehicle Functions

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper introduces SDVDiag, a modular platform that automates root-cause diagnosis for connected vehicle functions by combining distributed tracing, learned causal graphs, anomaly detection, and a weighted random walk; it is evaluated…

desk verdict A real integration of known components into a modular connected-vehicle diagnostic platform, but the single qualitative evaluation doesn't support the 'detected reliably' claim. read the letter →

arxiv 2507.19403 v1 pith:YPUANLHH submitted 2025-07-25 cs.SE cs.AIcs.DC

Matthias Wei{ss} , Falk Dettinger , Michael Weyrich This is my paper

classification cs.SEcs.AIcs.DC

keywords ConnectedVehicleRoot-causeAnalysisCausalityMiningAnomalyDetectionDependencyGraphDistributedTracingSoftware-definedRandomWalkRanking

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Connected and software-defined vehicles depend on services spread across the vehicle, edge nodes, and the cloud, so a single failure can hide behind a long chain of dependencies. SDVDiag is a platform that automates the diagnosis of such failures: it builds a dependency graph from live request traces, learns causal links between performance metrics, merges and prunes these graphs, and ranks the most probable root causes with a random walk. The paper's claim is that this pipeline detects injected faults reliably in a real 5G test-fleet deployment, demonstrated on a smart-charging scenario where one service's CPU spike triggers observable effects across the cluster. If this holds, engineers could go straight to a short list of suspected causes instead of manually navigating telemetry from vehicles and backend services.

What carries the argument

The load-bearing mechanism is the extended causal graph together with the first-order random walk. Dependency edges come from distributed traces recorded while requests flow through services; causal edges come from Amortized Causal Discovery, a deep-learning method that infers causal structure from time-series metrics, with edge weights representing causal strength. The two sources are merged by attaching metric nodes to their service instances, and causal edges are pruned whenever the corresponding services are not connected in the dependency graph, suppressing causalities that the live communication structure does not support. Given an incident snapshot with anomalies, the random walk starts at anomaly nodes, traverses graph edges with probability proportional to causal strength, and generates a ranking from accumulated visit counts. The platform's learning environment retrains the causal model after system changes and selects anomaly-detection models per time series.

What would settle it

Inject faults at many randomly chosen services in the same 5G test fleet, with known ground-truth root causes, and record where the true cause lands in the ranked list for each incident; if true causes often appear outside the top few positions, the claim that injected faults are detected reliably would be refuted.

Watch

Extended reading notes

Core claim

The paper's own claim is that a diagnosis platform combining two complementary graph views can identify root causes in distributed connected-vehicle functions: a dependency graph inferred from distributed tracing shows who talks to whom, while a causal graph learned from metric time series quantifies how one component's behavior affects another. SDVDiag merges these into an extended causal graph, prunes causal edges whose services are not actually related by a dependency, snapshots the graph when an incident is detected, and then runs a first-order random walk whose transition weights are the causal strengths. The walk's visit counts produce a ranked list of the most probable root causes. In the reported evaluation, a single charging-station service under increased CPU load produces cascading CPU effects across worker nodes, and SDVDiag's ranking isolates that service as the root cause, which the paper attributes to the accuracy of the causal discovery model and the graph pruning.

Load-bearing premise

The whole diagnosis rests on the learned causal graph being faithful: if the automatically inferred links between metric signals are wrong or missing, the random walk will rank the wrong component at the top, and the evaluation's single scenario does not test that risk widely.

Editorial extensions

If this is right

Engineers get an ordered list of likely causes per incident instead of manually tracing logs and metrics through vehicle, edge, and cloud services.
Because dependency graphs are rebuilt from live traces and causal models are retrained after system changes, the diagnosis reflects the current fleet state rather than a static architecture diagram.
The modular, runtime-exchangeable pipeline means new anomaly-detection and causal-discovery algorithms can be swapped in as they become available, without rebuilding the platform.
The same graph-based approach works across in-vehicle, edge, and cloud components, extending diagnostics beyond traditional on-board protocols to the backend services that connected vehicle functions depend on.
In larger systems with longer causal chains, the platform offers second-order random walks and fault tree analysis as alternatives to the default first-order walk.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Beyond the paper, the reliability claim would be more believable if tested across many injected faults: measuring how often the true root cause appears in the top-ranked positions across, say, dozens of random fault injections would quantify the platform's diagnosis accuracy.
Because the paper acknowledges that causal-model quality degrades quickly after minor system changes, an extension is to measure ranking accuracy immediately after a software update and to compare how much retraining latency worsens diagnosis.
The anomaly-detection UI shows early false positives that decline over time; a testable extension is to compare the adaptive model selector's false-positive rate and detection delay against a fixed, non-adaptive detector on the same live metrics.
The feedback loop and expert-knowledge integration are left for future work; if implemented, the platform could in principle handle failures that are not visible in metrics, such as semantic errors in service logic, by adding expert-supplied edges to the causal graph.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a genuine integration paper. The new thing is stitching OpenTelemetry-style tracing, ACD causal discovery, an anomaly-detection model selector, dependency-graph pruning, and random-walk ranking into one modular platform aimed at vehicle-edge-cloud diagnosis. That combination, deployed on a real 5G testbed with UGVs and Kubernetes, is new in the automotive software engineering literature, and the architecture is sensibly designed: separating data aggregation, graph building, learning, and incident analysis is the right call.

Credit where due: the authors clearly know the stack. They acknowledge ACD's fragility and are upfront that the human feedback loop is future work. The graph-pruning idea—using the dependency graph to cut causal edges between unrelated services—is a practical, sensible way to reduce the search space. The deployment itself, with OpenTelemetry, Kafka, Delta Lake, and a live 5G test fleet, is real engineering and not a toy simulation.

The soft spot is exactly where the stress-test note lands. The evaluation is one injected CPU fault, and the paper claims SDVDiag 'successfully isolates the root cause' without ever showing the random-walk ranking output. No ranked list, no hit@k, no MRR, no baseline, no repeated trials. The figure shows a dense anomalous graph; the text asserts the conclusion. That is not enough to support 'injected faults can be detected reliably' in the abstract. The self-adaptive loop that distinguishes the platform is also deferred, so the central selling point is aspirational rather than demonstrated. This is a proof-of-concept demo, not an evaluation.

The citation pattern looks fine; the self-citations are to the authors' earlier work on connected vehicle data loops and anomaly detection, which is relevant context. Nothing circular there.

Who is this for? People building diagnostic platforms for connected vehicles will get a useful blueprint. Researchers comparing RCA methods will find the evaluation too thin to use as a benchmark. It deserves a serious referee because the engineering problem is real and the platform is plausible, but the referee should demand either quantitative evidence or significantly softened claims. My recommendation: send it out, with a clear expectation of major revision. This is not a desk reject, but it needs more than polish.

Referee Report

4 major / 4 minor

Summary. The paper presents SDVDiag, a modular platform for automated diagnosis of connected vehicle functions. The platform combines distributed tracing, dependency and causal graph construction, adaptive anomaly detection, and random-walk-based root-cause ranking. The architecture is described across data aggregation, graph creation, a learning environment, and incident analysis, with an emphasis on runtime module exchange and continuous model updating. The evaluation deploys the platform in a 5G-connected test fleet with a smart charging application and reports a single scenario with an injected CPU fault, claiming that the platform reliably detects anomalies and isolates the root cause. The manuscript also identifies several limitations, notably the sensitivity of causal discovery models to system changes and the deferred design of human-feedback loops.

Significance. If the central claims were quantitatively supported, the paper would make a useful systems contribution by transferring observability and causality-mining techniques from generic microservice environments to the connected-vehicle domain, where vehicle, edge, and cloud components must be diagnosed jointly. The strengths are the real 5G test-fleet deployment, the modular service-based architecture that supports runtime model exchange, and the use of a modern causal discovery framework (ACD) rather than a purely static dependency model. However, the current evaluation is a single qualitative scenario, so the paper's significance depends on evidence that is not yet reported.

major comments (4)

[Section V] The central claim that SDVDiag 'successfully isolates the root cause of the incident' is not supported by any quantitative evidence. The first-order random-walk ranking described in Section IV-E is never reported: no ranked list, no position of the true cause, and no ranking metric such as hit@k or MRR. The statement in Fig. 5 that the graph is a 'dense network of causal relationships' makes it especially important to show that the ranking step, not merely the graph visualization, places the affected charging-station service at the top.
[Section V, Fig. 4] The claim that the AD Model Selector improves over time and that 'false positives decrease' is qualitative only. The paper gives no counts, no false-positive rate, no comparison against a static anomaly detector, and no repeated trials. Without these numbers, the statement that SDVDiag 'detects anomalies reliably' remains unverified.
[Sections IV-C, IV-D, IV-E] The correctness of the root-cause ranking depends on the faithfulness of the ACD-derived causal edges and on the pruning decisions, yet the paper itself admits in Section IV-D that 'even minor system changes can rapidly degrade the quality of results.' The evaluation contains no sensitivity analysis, no ablation of the pruning step, no variation of the injected fault, and no repeated runs. This is a load-bearing gap because a single successful-looking scenario cannot establish that the random-walk traversal over the pruned graph ranks true causes above non-causes in general.
[Sections IV-D and VI] The paper repeatedly describes continuous model adaptation as a platform capability, but the detailed feedback loops for the causal model encoder and the anomaly-detection model selector are explicitly deferred to future work. In the evaluation, the only adaptation evidence is the qualitative reduction of false positives in Fig. 4. The conclusion that the platform 'provides effective mechanisms for continuous model adaptation on live operational data' therefore goes beyond what is demonstrated.

minor comments (4)

[Abstract and Section I] There are several typographical errors: 'an 5G' should be 'a 5G', and 'Secton III' in the introduction should be 'Section III'.
[Section IV-E] The random-walk ranking algorithm is described only verbally; a formal definition of the edge-weight normalization and the visit-count aggregation would improve reproducibility.
[References] References [13] and [20] are the same paper title and author list with slightly different metadata; this duplication should be resolved to avoid confusion.
[Section IV-C, Fig. 2] The caption of Fig. 2 states that 'Causalities are pruned when there is no direct dependency between two nodes,' but the main text does not specify the threshold or criterion for pruning in enough detail for the result to be reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SDVDiag's pipeline combines external causal discovery, tracing, and graph traversal; the unsupported root-cause success claim is an evaluation gap, not a derivation that reduces to its inputs.

full rationale

The paper's derivation chain is architectural and empirical rather than formal: dependency graphs come from distributed traces, causal graphs come from the external Amortized Causal Discovery framework of Loewe et al. [22], and incident analysis is a weighted random-walk ranking over the combined, pruned graph. None of these steps is defined in terms of the target root-cause output. The pruning rules (Section IV-C: prune causal edges when services are not related in the dependency graph; Section IV-E: prune paths through services without anomalies) constrain the search space but do not force a particular ranking; the random walk then ranks by causal edge weights. The central claim in Section V that SDVDiag 'successfully isolates the root cause' is unsupported because no ranked list or quantitative ranking metric is reported, and the causal model robustness limitation is acknowledged in Section IV-D. However, missing evidence is a correctness/validation weakness, not circularity: there is no equation, fitted parameter, or self-citation chain that makes the conclusion equivalent to an input by construction. Self-citations [1]-[5], [23] provide contextual background and are not load-bearing for the diagnosis claim; the load-bearing causal discovery citation [22] is external prior work. Therefore the derivation is self-contained with respect to circularity concerns.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The platform's correctness rests on several unverified domain assumptions about trace completeness, causal graph fidelity, pruning safety, and the meaning of random-walk counts. No external benchmark or quantitative ground-truth comparison is provided. The two learned models are fitted to operational data with undisclosed details, so they function as implicit free parameters in the demonstration.

free parameters (2)

ACD causal model weights = not disclosed
The causal graph that drives root-cause ranking is produced by a trained ACD encoder; training data, architecture, and hyperparameters are not reported in Sections IV-C and IV-D.
Anomaly detection model selector policy = not disclosed
An agent trained on historical labeled data selects anomaly detectors per metric; no training details or evaluation of selector accuracy are given in Section IV-D.

assumptions (4)

domain assumption Complete dependency graph recoverable from distributed traces alone.
Section IV-C derives dependency edges from live communication flows; untraced or asynchronous communication would create missing edges and mislead pruning.
domain assumption ACD-inferred causal edges are faithful enough for root-cause ranking.
Section IV-C relies on ACD to quantify causal strength, while Section IV-D notes that ACD quality can degrade quickly with minor system changes.
domain assumption Pruning services without anomalies preserves the true root cause.
Section IV-E prunes graph paths involving services without detected anomalies, assuming they are unlikely to contribute; if anomalies are missed, the root cause could be pruned.
domain assumption First-order random walk visit counts approximate root-cause probability.
Section IV-E ranks causes by visit counts over causally weighted edges, but no justification is given that this approximates causal likelihood.

how reviews work

0 comments

Cite this review

Pith. "Pith review of SDVDiag: A Modular Platform for the Diagnosis of Connected Vehicle Functions." pith.science (2026). https://pith.science/paper/YPUANLHH

@misc{pith2026250719403,
  author       = {Pith},
  title        = {Pith review of: SDVDiag: A Modular Platform for the Diagnosis of Connected Vehicle Functions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YPUANLHH}},
  note         = {Machine review of arXiv:2507.19403}
}

read the original abstract

Connected and software-defined vehicles promise to offer a broad range of services and advanced functions to customers, aiming to increase passenger comfort and support autonomous driving capabilities. Due to the high reliability and availability requirements of connected vehicles, it is crucial to resolve any occurring failures quickly. To achieve this however, a complex cloud/edge architecture with a mesh of dependencies must be navigated to diagnose the responsible root cause. As such, manual analyses become unfeasible since they would significantly delay the troubleshooting. To address this challenge, this paper presents SDVDiag, an extensible platform for the automated diagnosis of connected vehicle functions. The platform enables the creation of pipelines that cover all steps from initial data collection to the tracing of potential root causes. In addition, SDVDiag supports self-adaptive behavior by the ability to exchange modules at runtime. Dependencies between functions are detected and continuously updated, resulting in a dynamic graph view of the system. In addition, vital system metrics are monitored for anomalies. Whenever an incident is investigated, a snapshot of the graph is taken and augmented by relevant anomalies. Finally, the analysis is performed by traversing the graph and creating a ranking of the most likely causes. To evaluate the platform, it is deployed inside an 5G test fleet environment for connected vehicle functions. The results show that injected faults can be detected reliably. As such, the platform offers the potential to gain new insights and reduce downtime by identifying problems and their causes at an early stage.

Figures

Figures reproduced from arXiv: 2507.19403 by the authors.

**Figure 1.** Conceptual Overview of SDVDiag. The platform can roughly be divided in subsystems for the generation of graphs, the analysis of these graphs for [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** Extended causality graph of SDVDiag. Information about service [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** UI of SDVDiag for Anomaly Detection. Yellow dots symbolize [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Figures from the paper (2 more)

**Figure 3.** Figure 3: System overview consisting of Unmanned Ground Vehicle with [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]

**Figure 5.** Figure 5: Anomalous causality graph for the charging station test scenario. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

24 extracted references · 18 canonical work pages

[1]

Connected vehicle: Ontology, taxonomy and use cases,

D. Baumann, M. Sommer, F. Dettinger, T. R ¨osch, M. Weyrich, and E. Sax, “Connected vehicle: Ontology, taxonomy and use cases,” in 2024 IEEE International Systems Conference (SysCon) , 2024, pp. 1–6

work page 2024
[2]

Contin- uous analysis and optimization of vehicle software updates using the intelligent digital twin,

M. Weiß, M. M ¨uller, F. Dettinger, N. Jazdi, and M. Weyrich, “Contin- uous analysis and optimization of vehicle software updates using the intelligent digital twin,” in 2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETF A) . IEEE, 2023, pp. 1–7

work page 2023
[3]

A survey about self- adaptive anomaly-detection in software-defined systems,

M. Weiß, S. Thich, M. Artelt, and M. Weyrich, “A survey about self- adaptive anomaly-detection in software-defined systems,” in IECON 2024 - 50th Annual Conference of the IEEE Industrial Electronics Society, 2024, pp. 1–4

2024
[4]

Devops als enabler der kontinuierlichen funktionsverbesserung und automatisierten update- analyse in software-definierten systemen,

M. Weiß, F. Dettinger, N. Jazdi, and M. Weyrich, “Devops als enabler der kontinuierlichen funktionsverbesserung und automatisierten update- analyse in software-definierten systemen,” in Automation 2023 , 2023

work page 2023
[5]

Future use cases for vehicular communication based on connected functions,

F. Dettinger, M. Weiß, and M. Weyrich, “Future use cases for vehicular communication based on connected functions,” in 2024 IEEE 100th V ehicular Technology Conference (VTC2024-Fall), 2024, pp. 1–5

work page 2024
[6]

Mobile cloud computing for computation offloading: Issues and challenges,

K. Akherfi, M. Gerndt, and H. Harroud, “Mobile cloud computing for computation offloading: Issues and challenges,” Applied computing and informatics, vol. 14, no. 1, pp. 1–16, 2018

work page 2018
[7]

A survey on mobile edge computing: The communication perspective,

Y . Mao, C. You, J. Zhang, K. Huang, and K. B. Letaief, “A survey on mobile edge computing: The communication perspective,” IEEE communications surveys & tutorials , vol. 19, no. 4, pp. 2322–2358, 2017

work page 2017
[8]

A survey and taxonomy on task offloading for edge-cloud computing,

B. Wang, C. Wang, W. Huang, Y . Song, and X. Qin, “A survey and taxonomy on task offloading for edge-cloud computing,” IEEE Access, vol. 8, pp. 186 080–186 101, 2020

work page 2020

Show all 24 references

[9]

6g white paper on edge intelligence,

E. Peltonen, M. Bennis, M. Capobianco, M. Debbah, A. Ding, F. Gil- Casti˜neira, M. Jurmu, T. Karvonen, M. Kelanti, A. Kliks, T. Lepp ¨anen, L. Lov´en, T. Mikkonen, A. Rao, S. Samarakoon, K. Sepp¨anen, P. Sroka, S. Tarkoma, and T. Yang, “6g white paper on edge intelligence,” 20...

2020 arXiv
[10]

Distributed systems observability,

C. Sridharan, “Distributed systems observability,” in V elocity Confer- ence. O’Reilly Media, 2018

2018
[11]

On ob- servability and monitoring of distributed systems: An industry inter- view study,

S. Niedermaier, F. Koetter, A. Freymann, and S. Wagner, “On ob- servability and monitoring of distributed systems: An industry inter- view study,” in Proceedings of the 15th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE) . SciTePress...

2019
[12]

Dapper, a large-scale dis- tributed systems tracing infrastructure,

B. H. Sigelman, L. A. Barroso, M. Burrows, P. Stephenson, M. Plakal, D. Beaver, S. Jaspan, and C. Shanbhag, “Dapper, a large-scale dis- tributed systems tracing infrastructure,” Google Research Technical Report, 2010

2010
[13]

Enjoy your observability: an industrial survey of microservice tracing and analysis,

B. Li, X. Peng, Q. Xiang, H. Wang, T. Xie, J. Sun, and X. Liu, “Enjoy your observability: an industrial survey of microservice tracing and analysis,” Empirical Software Engineering , vol. 27, no. 1, pp. 1–42, 2022

2022
[14]

Anomaly detection: A survey,

V . Chandola, A. Banerjee, and V . Kumar, “Anomaly detection: A survey,” ACM Computing Surveys (CSUR) , vol. 41, no. 3, pp. 1–58, 2009

2009
[15]

A unifying review of deep and shallow anomaly detection,

L. Ruff, J. R. Kauffmann, R. A. Vandermeulen, G. Montavon, W. Samek, M. Kloft, T. G. Dietterich, and K.-R. M ¨uller, “A unifying review of deep and shallow anomaly detection,” Proceedings of the IEEE, vol. 109, no. 5, pp. 756–795, 2021

2021
[16]

Road vehicles — uni- fied diagnostic services (uds) — part 1: Application layer,

International Organization for Standardization, “Road vehicles — uni- fied diagnostic services (uds) — part 1: Application layer,” International Organization for Standardization, Geneva, Switzerland, Tech. Rep. ISO 14229-1:2020, 2020

2020
[17]

G. Team. (2023) What is obdii? history of on-board diagnostics (obd). Accessed: 2025-05-05. [Online]. Available: https://www.geotab.com/blog/obd-ii/

2023
[18]

Serial control and communications heavy-duty vehicle network,

SAE, “Serial control and communications heavy-duty vehicle network,” SAE International, Warrendale, PA, USA, Tech. Rep. J1939 202306, June 2023, revised June 2023, originally issued April 2000. [Online]. Available: https://www.sae.org/standards/content/j1939 202306/

2023
[19]

Service-oriented vehicle diagnostics sovd),

ASAM, “Service-oriented vehicle diagnostics sovd),” Association for Standardisation of Automation and Measuring Systems (ASAM), Ingol- stadt, Germany, Tech. Rep. Version 1.0.0, Jun. 2022, accessed: 2025- 05-04. [Online]. Available: https://www.asam.net/standards/detail/sovd/

2022
[20]

Enjoy your observability: an industrial survey of microservice tracing and analysis,

B. Li, X. Peng, Q. Xiang, H. Wang, T. Xie, J. Sun, and X. Liu, “Enjoy your observability: an industrial survey of microservice tracing and analysis,” Empirical Software Engineering , vol. 27, no. 1, p. 25,
[21]

A comprehensive survey on root cause analysis in (micro) services: Methodologies, challenges, and trends,

T. Wang and G. Qi, “A comprehensive survey on root cause analysis in (micro) services: Methodologies, challenges, and trends,” 2024. [Online]. Available: https://arxiv.org/abs/2408.00803

2024 arXiv
[22]

Amortized causal discovery: Learning to infer causal graphs from time-series data,

S. L ¨owe, D. Madras, R. Zemel, and M. Welling, “Amortized causal discovery: Learning to infer causal graphs from time-series data,” in Conference on Causal Learning and Reasoning . PMLR, 2022, pp. 509–525

2022
[23]

Simulating cloud environments of connected vehicles for anomaly detection,

M. Weiß, J. St ¨umpfle, F. Dettinger, N. Jazdi, and M. Weyrich, “Simulating cloud environments of connected vehicles for anomaly detection,” in SAE Technical Paper Series , ser. STUT. SAE International, Jul. 2024. [Online]. Available: http://dx.doi.org/10.4271/2024-01-2996

2024 doi
[2021]

Available: https://doi.org/10.1007/s10664-021-10063-9

[Online]. Available: https://doi.org/10.1007/s10664-021-10063-9

doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Connected vehicle: Ontology, taxonomy and use cases,

D. Baumann, M. Sommer, F. Dettinger, T. R ¨osch, M. Weyrich, and E. Sax, “Connected vehicle: Ontology, taxonomy and use cases,” in 2024 IEEE International Systems Conference (SysCon) , 2024, pp. 1–6

work page 2024

[2] [2]

Contin- uous analysis and optimization of vehicle software updates using the intelligent digital twin,

M. Weiß, M. M ¨uller, F. Dettinger, N. Jazdi, and M. Weyrich, “Contin- uous analysis and optimization of vehicle software updates using the intelligent digital twin,” in 2023 IEEE 28th International Conference on Emerging Technologies and Factory Automation (ETF A) . IEEE, 2023, pp. 1–7

work page 2023

[3] [3]

A survey about self- adaptive anomaly-detection in software-defined systems,

M. Weiß, S. Thich, M. Artelt, and M. Weyrich, “A survey about self- adaptive anomaly-detection in software-defined systems,” in IECON 2024 - 50th Annual Conference of the IEEE Industrial Electronics Society, 2024, pp. 1–4

2024

[4] [4]

Devops als enabler der kontinuierlichen funktionsverbesserung und automatisierten update- analyse in software-definierten systemen,

M. Weiß, F. Dettinger, N. Jazdi, and M. Weyrich, “Devops als enabler der kontinuierlichen funktionsverbesserung und automatisierten update- analyse in software-definierten systemen,” in Automation 2023 , 2023

work page 2023

[5] [5]

Future use cases for vehicular communication based on connected functions,

F. Dettinger, M. Weiß, and M. Weyrich, “Future use cases for vehicular communication based on connected functions,” in 2024 IEEE 100th V ehicular Technology Conference (VTC2024-Fall), 2024, pp. 1–5

work page 2024

[6] [6]

Mobile cloud computing for computation offloading: Issues and challenges,

K. Akherfi, M. Gerndt, and H. Harroud, “Mobile cloud computing for computation offloading: Issues and challenges,” Applied computing and informatics, vol. 14, no. 1, pp. 1–16, 2018

work page 2018

[7] [7]

A survey on mobile edge computing: The communication perspective,

Y . Mao, C. You, J. Zhang, K. Huang, and K. B. Letaief, “A survey on mobile edge computing: The communication perspective,” IEEE communications surveys & tutorials , vol. 19, no. 4, pp. 2322–2358, 2017

work page 2017

[8] [8]

A survey and taxonomy on task offloading for edge-cloud computing,

B. Wang, C. Wang, W. Huang, Y . Song, and X. Qin, “A survey and taxonomy on task offloading for edge-cloud computing,” IEEE Access, vol. 8, pp. 186 080–186 101, 2020

work page 2020