Γ-Cantilever Mechanism-Informed BP Neural Network Compensation for Gap Measurement of an Integrated Knee Distractor in Total Knee Arthroplasty

In total knee arthroplasty (TKA), intraoperative soft-tissue balance and joint gap accuracy directly affect postoperative joint stability, functional recovery and long-term implant survivorship, but current clinical practice still lacks stable, quantitative measurement tools with sufficient precision. This study developed an integrated knee distractor with integrated force feedback and gap detection; the hardware design reserves two independent actuation channels, whereas all experiments reported in this study were conducted in single-channel mode. A mechanism-informed nonlinear error compensation method was proposed to reduce load-dependent gap-measurement error. First, the load-bearing structure was abstracted as a Γ-shaped cantilever beam model under small-deformation and linear-elastic assumptions, and mechanical analysis was conducted to clarify the formation mechanism of gap measurement error induced by load-induced structural deformation and assembly inaccuracies. Guided by this mechanistic analysis, a dual-input BP neural network was constructed with displacement-sensor output and measured force as inputs and the actual distraction gap as output; the mechanical model informed the selection of these two inputs (mechanism-informed input selection) rather than imposing deformation equations as training constraints. The model was trained and evaluated using 200 averaged calibration samples, constructed from a structured grid of 40 displacement settings (0.5–20 mm at 0.5 mm intervals) crossed with five nominal force levels (20, 40, 60, 80, and 100 N), split at the averaged-sample level into 140 training, 20 validation, and 40 test samples. On the independent test set, the proposed method achieved a root mean square error (RMSE) of 0.1123 mm, a mean absolute error (MAE) of 0.0893 mm, and a maximum absolute error (MaxAE) of 0.3189 mm, with all 40 test samples having absolute errors below the predefined engineering accuracy target of ±0.5 mm under the specified bench-top conditions. This study provides a promising engineering approach for reducing load-dependent gap-measurement error under controlled bench-top conditions; validation on anatomically and biomechanically realistic systems is required before any claim of intraoperative suitability.

Authors

Institutions

Publication Details

Journal
Sensors
Published
2026-11-13
DOI
https://doi.org/10.3390/s26227169
Primary Topic
Total Knee Arthroplasty Outcomes
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Γ-Cantilever Mechanism-Informed BP Neural Network Compensation for Gap Measurement of an Integrated Knee Distractor in Total Knee Arthroplasty

Enkui Zhang, He Xu, Tailin Han, Mingchi Ju et al.
2 links
Sensors
Total Knee Arthroplasty Outcomes
article

Γ-Cantilever Mechanism-Informed BP Neural Network Compensation for Gap Measurement of an Integrated Knee Distractor in Total Knee Arthroplasty

Enkui Zhang, He Xu, Tailin Han, Mingchi Ju, Xuan Liu, Jun Hu, Guanyu Chen
article en

Abstract

In total knee arthroplasty (TKA), intraoperative soft-tissue balance and joint gap accuracy directly affect postoperative joint stability, functional recovery and long-term implant survivorship, but current clinical practice still lacks stable, quantitative measurement tools with sufficient precision. This study developed an integrated knee distractor with integrated force feedback and gap detection; the hardware design reserves two independent actuation channels, whereas all experiments reported in this study were conducted in single-channel mode. A mechanism-informed nonlinear error compensation method was proposed to reduce load-dependent gap-measurement error. First, the load-bearing structure was abstracted as a Γ-shaped cantilever beam model under small-deformation and linear-elastic assumptions, and mechanical analysis was conducted to clarify the formation mechanism of gap measurement error induced by load-induced structural deformation and assembly inaccuracies. Guided by this mechanistic analysis, a dual-input BP neural network was constructed with displacement-sensor output and measured force as inputs and the actual distraction gap as output; the mechanical model informed the selection of these two inputs (mechanism-informed input selection) rather than imposing deformation equations as training constraints. The model was trained and evaluated using 200 averaged calibration samples, constructed from a structured grid of 40 displacement settings (0.5–20 mm at 0.5 mm intervals) crossed with five nominal force levels (20, 40, 60, 80, and 100 N), split at the averaged-sample level into 140 training, 20 validation, and 40 test samples. On the independent test set, the proposed method achieved a root mean square error (RMSE) of 0.1123 mm, a mean absolute error (MAE) of 0.0893 mm, and a maximum absolute error (MaxAE) of 0.3189 mm, with all 40 test samples having absolute errors below the predefined engineering accuracy target of ±0.5 mm under the specified bench-top conditions. This study provides a promising engineering approach for reducing load-dependent gap-measurement error under controlled bench-top conditions; validation on anatomically and biomechanically realistic systems is required before any claim of intraoperative suitability.

SensorsVol. 26(22)
Changchun University of Science and Technology (CN)
Good health and well-being
Total Knee Arthroplasty Outcomes
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon