CoChat

A Review of Embedded Machine Learning Based on Hardware, Application, and Sensing Scheme

Open access
Biglari, Amin, Tang, Wei · 2023 · Sensors

The DOI matches its Crossref record; 181 of the 189 references it cites were found in a registry or an index; 2 of the 59 sources its claims cite were checked for support.

Abstract

Machine learning is an expanding field with an ever-increasing role in everyday life, with its utility in the industrial, agricultural, and medical sectors being undeniable. Recently, this utility has come in the form of machine learning implementation on embedded system devices. While there have been steady advances in the performance, memory, and power consumption of embedded devices, most machine learning algorithms still have a very high power consumption and computational demand, making the implementation of embedded machine learning somewhat difficult. However, different devices can be implemented for different applications based on their overall processing power and performance. This paper presents an overview of several different implementations of machine learning on embedded systems divided by their specific device, application, specific machine learning algorithm, and sensors. We will mainly focus on NVIDIA Jetson and Raspberry Pi devices with a few different less utilized embedded computers, as well as which of these devices were more commonly used for specific applications in different fields. We will also briefly analyze the specific ML models most commonly implemented on the devices and the specific sensors that were used to gather input from the field. All of the papers included in this review were selected using Google Scholar and published papers in the IEEExplore database. The selection criterion for these papers was the usage of embedded computing systems in either a theoretical study or practical implementation of machine learning models. The papers needed to have provided either one or, preferably, all of the following results in their studies-the overall accuracy of the models on the system, the overall power consumption of the embedded machine learning system, and the inference time of their models on the embedded system. Embedded machine learning is experiencing an explosion in both scale and scope, both due to advances in system performance and machine learning models, as well as greater affordability and accessibility of both. Improvements are noted in quality, power usage, and effectiveness.

Overview

CoChat Summary

TL;DR

This review surveys embedded machine learning implementations across hardware platforms, applications, and sensing schemes, focusing mainly on NVIDIA Jetson and Raspberry Pi devices with several less-used boards. It compiles reported accuracy, inference speed, and power consumption to compare which embedded systems suit which applications.

Key findings

  • The review centers on NVIDIA Jetson and Raspberry Pi, with other boards including Google Coral, ASUS Tinker, ODROID-XU4, and Banana Pi; reported hardware examples include Jetson Nano at 5–10 W and USD 300–500, Coral Dev Board at 6–10 W and USD 130, Raspberry Pi 3 Model B at 1.5 W and about USD 40, Raspberry Pi 4 Model B at 4 W and USD 40–80, and Jetson AGX Xavier at 30 W and about USD 4000.
  • Applications are grouped into autonomous driving, security and safety, healthcare, drones, and agriculture; priorities differ, with autonomous driving emphasizing performance speed, drones emphasizing accuracy, and agriculture emphasizing detection accuracy and energy consumption over speed.
  • Sensors covered include RGB, infrared/thermal, depth, 360-degree cameras, radar, LiDAR, microphones, body motion sensors, electrocardiograms, and electroencephalograms; most reviewed papers used some form of computer vision.
  • The review compares systems by performance and cost, using the mean of all final results to assess overall performance, with results presented in Tables 2–13 and Figures 2–11.

Methods

  • Review/compilation of referenced research papers; sources were selected using Google Scholar and published papers in the IEEE Xplore database, with inclusion requiring use of embedded computing systems in a theoretical study or practical implementation of machine learning models.
  • Included papers needed to report one or preferably all of: model accuracy on the system, overall power consumption of the embedded machine learning system, and inference time on the embedded system.
  • Sources were divided by device, application, and sensing scheme; systems were compared by performance and cost, with performance assessed differently depending on the task and the mean of final results used for overall performance.

Limitations

  • The paper states that not all sources made active use of a sensor, with some mainly exploring theoretical implementation; non-vision-related applications had insufficient numbers for a proper basis-by-basis comparison and were displayed in Table 10.
  • The paper states that most embedded machine learning optimization papers did not have sensing schemes, and most benchmark/review papers in that category also did not include sensing schemes.
  • Evident from the text: the performance analysis method differed from source to source and depended heavily on the specific application and sensory system, limiting direct comparison across studies.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/s23042131

Cites

181 of 181 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 2 Average inference time in agricultural computer vision for devices used in this application.
Figure 2 Average inference time in agricultural computer vision for devices used in this application.CC BY · PubMed Central
Figure 3 Average Inference time in facial recognition for devices used in this application.
Figure 3 Average Inference time in facial recognition for devices used in this application.CC BY · PubMed Central
Figure 4 Avg. inference time in depth estimation for devices used in this application.
Figure 4 Avg. inference time in depth estimation for devices used in this application.CC BY · PubMed Central
Figure 5 Average inference time in autonomous vehicle obstacle recognition in devices used in this application.
Figure 5 Average inference time in autonomous vehicle obstacle recognition in devices used in this application.CC BY · PubMed Central
Figure 6 Average inference time in medicine and disability assistance in devices used in these applications.
Figure 6 Average inference time in medicine and disability assistance in devices used in these applications.CC BY · PubMed Central
Figure 7 Average inference time in safety and security in devices used in these applications.
Figure 7 Average inference time in safety and security in devices used in these applications.CC BY · PubMed Central
Figure 8 Average inference time in devices used in city management applications.
Figure 8 Average inference time in devices used in city management applications.CC BY · PubMed Central
Figure 9 Average inference time in embedded computer vision devices.
Figure 9 Average inference time in embedded computer vision devices.CC BY · PubMed Central
Figure 10 Average inference time in devices used for testing model optimization methods.
Figure 10 Average inference time in devices used for testing model optimization methods.CC BY · PubMed Central
Figure 11 Average inference time in devices covered in referenced benchmark papers.
Figure 11 Average inference time in devices covered in referenced benchmark papers.CC BY · PubMed Central
Figure 1 Paper Layout Showing the Distribution of Subjects Covered in the Review.
Figure 1 Paper Layout Showing the Distribution of Subjects Covered in the Review.CC BY · PubMed Central

Verification

Claims the paper makes

1 Valid · 12 Not valid · 30 Not checked · 34 Own results and interpretations

8 more claims

Create a free account to see all 77 claims and to verify the 37 references still open.

See all claims — free account
  • It may state two separate claims as one. 39
  • Probably not a substantial claim. 52

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 181 references

Cited by

See all 86

Metadata

Authors
Biglari, A., & Tang, W.
Affiliations
New Mexico State University (US)
Corresponding author
Tang, Wei
Venue
Sensors
Publisher
MDPI AG
Published
2023-02-14
Volume
23
Issue
4
Pages
2131
ISSN
1424-8220 (ISSN-L)
OpenAlex
W4320730661
Cited by
86
References
189
Citations by year
2026: 25 · 2025: 36 · 2024: 19 · 2023: 6
FWCI
2.23
Type
journal article
Language
en
Open access
Locations
  • Sensors (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Science Foundation (ECCS-1652944, ECCS-2015573)
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
  • Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
  • Physical Sciences › Engineering › Control and Systems Engineering › Embedded Systems and FPGA Design
Keywords
embedded machine learning, embedded systems, NVIDIA Jetson, Raspberry Pi, power consumption, inference latency, sensor input, machine learning applications, edge computing, low-power devices

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A Review of Embedded Machine Learning Based on Hardware, Application, and Sensing Scheme. https://app.cochat.ai/papers/10.3390/s23042131
The paper
Biglari, A., & Tang, W. (2023). A Review of Embedded Machine Learning Based on Hardware, Application, and Sensing Scheme. Sensors. https://doi.org/10.3390/s23042131

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.