PCA-Driven Adaptive Sensor Triage for Edge AI Inference

Ankit Hemant Lade; Sai Krishna Jasti; Nikhil Sinha; Indar Kumar; Akanksha Tiwari

cs.LGApr 6, 2026

PCA-Driven Adaptive Sensor Triage for Edge AI Inference

Authors: Ankit Hemant Lade, Sai Krishna Jasti, Nikhil Sinha, Indar Kumar, Akanksha Tiwari

Organizations: Independent Researchers

Abstract

Multi-channel sensor networks in industrial IoT often exceed available bandwidth. We propose PCA-Triage, a streaming algorithm that converts incremental PCA loadings into proportional per-channel sampling rates under a bandwidth budget. PCA-Triage runs in O(wdk) time with zero trainable parameters (0.67 ms per decision). We evaluate on 7 benchmarks (8--82 channels) against 9 baselines. PCA-Triage is the best unsupervised method on 3 of 6 datasets at 50% bandwidth, winning 5 of 6 against every baseline with large effect sizes (r = 0.71--0.91). On TEP, it achieves F1 = 0.961 +/- 0.001 -- within 0.1% of full-data performance -- while maintaining F1 > 0.90 at 30% budget. Targeted extensions push F1 to 0.970. The algorithm is robust to packet loss and sensor noise (3.7--4.8% degradation under combined worst-case).

Figures & tables

Fig. 1: TEP sensor correlation matrix during fault-free operation. Highly correlated clusters (dark red/blue blocks) represent redundancy that PCA-Triage exploits for efficient bandwidth allocation.

	Static/Batch	Streaming/Adaptive
Fixed Rules	Uniform, SoD	DSRA, AIMD
Data-Driven	Batch PCA, Offline FS	PCA-Triage (Ours)

TABLE I: Positioning matrix. PCA-Triage combines streaming operation with data-driven proportional allocation (bottom-right cell). Prior work [ 29 , 30 ] explores correlation-aware sensor selection but in batch or binary settings.

Fig. 2: PCA-Triage system architecture. Sensor data passes through a sliding window into IncrementalPCA, which extracts loadings $\mathbf{V}$ and singular values $\boldsymbol{\sigma}$ . Importance scores $s_{j}$ are computed from weighted loadings and fed to the rate allocator, which distributes the bandwidth budget $B$ across channels. Time: $O(wdk)$ ; Memory: $O(wd+kd)$ ; Trainable parameters: 0.

Dataset	Domain	Ch.	Samples	Source
TEP	Chemical process	52	250K	[ 3 ]
SMD	Server machines	38	388K	[ 33 ]
MSL	Spacecraft telemetry	55	132K	[ 34 ]
PSM	Server metrics	25	220K	[ 35 ]
HAI	Industrial control	82	259K	[ 36 ]
SKAB	Water circulation	8	47K	Skoltech

TABLE II: Datasets used in evaluation.

\dagger

= synthetic stand-in calibrated to match published dataset properties (real data requires institutional access agreement).

Fig. 3: TEP fault-free operation: 6 representative sensor channels showing diverse signal characteristics—smooth trends (Reactor Pressure), noisy oscillations (Separator Level), and step-like control signals (Stripper Steam Valve).

Fig. 4: Pareto curves: F1 vs. bandwidth budget across 6 datasets (5-seed average). PCA-Triage (blue) dominates on high-channel datasets with rich correlation structure (TEP, SMD, MSL). Error bands show $\pm 1$ standard deviation.

$B$	PCA-T	Var	Thr	Uni	R.D.
10%	.908	.909	.914	.920	.460
20%	.915	.912	.917	.918	.591
30%	.924	.917	.920	.918	.675
40%	.943	.927	.935	.919	.732
50%	.961	.948	.958	.924	.788
60%	.963	.959	.966	.927	.820

TABLE III: TEP F1 across bandwidth levels (3-seed mean, base PCA scoring). Bold = best unsupervised per row. PCA-Triage wins at

B\in\{30\%,40\%,50\%\}

—the operationally critical range.

Method	TEP	SMD	PSM	MSL	HAI	SKAB
PCA-Triage	$\mathbf{.961\pm.001}$	$\mathbf{.982\pm.001}$	$.959\pm.001$	$\mathbf{.921\pm.004}$	$1.000\pm.000$	$.583\pm.013$
Variance	$.948\pm.001$	$.977\pm.000$	$.903\pm.000$	$.917\pm.001$	$\mathbf{1.000\pm.000}$	$.577\pm.024$
Threshold	$.958\pm.004$	$.981\pm.000$	$.925\pm.001$	$.913\pm.001$	$1.000\pm.000$	$.582\pm.012$
Uniform	$.924\pm.002$	$.967\pm.000$	$.897\pm.001$	$.912\pm.001$	$.998\pm.000$	$\mathbf{.586\pm.009}$
Random Dropout	$.788\pm.003$	$.980\pm.000$	$\mathbf{.962\pm.000}^{\dagger}$	$.914\pm.000$	$1.000\pm.000$	$.524\pm.009$
Mutual Info †	$.914\pm.003$	$.986\pm.000^{\dagger}$	$.996\pm.000^{\dagger}$	$.930\pm.001^{\dagger}$	$1.000\pm.000$	$.588\pm.008$

TABLE IV: F1 at 50% bandwidth (mean

\pm

std, 5 seeds, RF

n{=}100

, forward-fill, no per-dataset tuning). Bold = best unsupervised per column. † = supervised or outperforms PCA-Triage. Cross-dataset Wilcoxon tests reported in Sec. VII .

Fig. 5: TEP fault detection F1 at 50% bandwidth. PCA-Triage (0.961) within 0.1% of full-data performance (0.962) and outperforms all unsupervised baselines with zero trainable parameters.

Dataset	Classifier	PCA-T	Uniform	Variance	Full
TEP	RF	.962	.920	.946	.961
	SVM	.960	.902	.933	.963
	KNN	.903	.881	.903	.926
SMD	RF	.981	.968	.976	.994
	SVM	.964	.956	.963	.972
	KNN	.973	.964	.971	.990

TABLE V: F1 at 50% bandwidth across classifiers. PCA-Triage consistently outperforms Uniform and Variance regardless of classifier choice.

Fig. 6: Channel importance adaptation under three TEP faults ( $\lambda=0.85$ , $w=25$ ). Top: Fault 1 (A/C Feed Ratio)—importance shifts to flow-related sensors. Middle: Fault 4 (Reactor CW Temp)—temperature sensors gain importance. Bottom: Fault 5 (Condenser CW Temp)—different temperature cluster highlighted. Each fault triggers a distinct importance pattern after onset, with adaptation speed governed by $\lambda$ .

Fig. 7: Detailed view: channel importance over time for TEP Fault 1 (A/C Feed Ratio disturbance). Top: full heatmap showing all 52 channels. Bottom: top-5 channel importance traces. The fault onset (dashed line) triggers a clear shift in the importance distribution.

Fig. 8: Allocated sampling rates over time for TEP Fault 1 at 50% budget. After fault onset (dashed line), bandwidth shifts toward fault-relevant channels (darker blue = higher rate).

Fault	PCA-Triage	Variance	Threshold
IDV(1): A/C Feed	19	0	19
IDV(2): B Composition	19	19	19
IDV(4): Reactor CW	19	19	19
IDV(5): Condenser CW	19	19	19

TABLE VI: Reaction time (windows until importance shift

>20\%

) for three TEP faults across triage methods (

\lambda=0.85

Fig. 9: Reaction time vs. forgetting factor $\lambda$ for four TEP faults. Lower $\lambda$ yields faster reaction (0–3 windows at $\lambda=0.80$ ) but at the cost of noisier importance estimates. $\lambda=0.85$ offers a practical balance.

Fig. 10: Ablation studies on TEP at 50% bandwidth. (a) Effect of $k$ : robust for $k\in[3,10]$ , degrades above 15 as noise components are included. (b) Effect of $w$ : stable across $w\in[50,500]$ . (c) Effect of $\lambda$ : $\lambda=1.0$ gives best F1 but slowest reaction. (d) Score formula comparison: weighted and unweighted loadings outperform reconstruction-based scoring.

Param	Value	F1 Mean	F1 Std
$k$	3	0.962	0.003
	5	0.962	0.003
	8	0.962	0.002
	10	0.962	0.001
	15	0.955	0.002
	20	0.939	0.004

TABLE VII: Ablation results on TEP (50% BW, 5 seeds). Each row varies one parameter while holding others at defaults (

k{=}10

w{=}50

\lambda{=}1.0

Configuration	F1	$\pm$ Std	$\Delta$ F1
Full PCA-Triage	0.961	0.002	—
No data-driven (Uniform)	0.924	0.002	$-$ 3.93%
No PCA (Variance)	0.949	0.001	$-$ 1.30%
No smoothing ( $\lambda{=}0.001$ )	0.954	0.002	$-$ 0.74%
Aggressive smooth ( $\lambda{=}0.5$ )	0.954	0.002	$-$ 0.77%
No proportional (Threshold)	0.958	0.004	$-$ 0.31%

TABLE VIII: Component contribution analysis on TEP (50% BW, 5 seeds, base PCA scoring with forward-fill). Each row disables one component.

\Delta

= change from base PCA-Triage (0.961). With targeted extensions (hybrid scoring, linear interpolation, sharpening), F1 reaches 0.970.

Budget	Fwd-Fill	Linear	Zero
10%	.910	.928 (+1.8%)	.433
30%	.925	.944 (+1.9%)	.574
50%	.962	.968 (+0.7%)	.757
70%	.962	.966 (+0.4%)	.930

TABLE IX: Reconstruction method comparison on TEP (3 seeds). Linear interpolation outperforms forward-fill at all budgets, with the largest gain at low bandwidth.

Method	ms (edge)	ms (laptop)	Peak MB	Coeff. ‡
PCA-Triage	0.67	1.46	8.5	520
Uniform	0.06	0.07	24.8	0
Threshold	0.16	0.75	8.1	0
Variance	0.15	0.70	8.1	0
Random Dropout	0.06	0.12	8.1	0
Random Proj.	0.18	0.70	8.2	9,600

TABLE X: Computational profile (TEP,

d{=}52

w{=}100

). Edge = single-threaded simulation; Laptop = multi-threaded measurement. ‡ Coefficients = computed model values (

kd

for PCA), not trainable parameters.

Fig. 11: Compute cost per triage decision (left) and peak memory usage (right). PCA-Triage (1.46 ms laptop / 0.67 ms edge) is well under the 5 ms edge target. Mutual Info is an order of magnitude slower.

Fig. 12: Scalability: compute time vs. number of channels (log-log). PCA-Triage ( $O(wdk)$ , blue) stays under the 5 ms edge target up to 500 channels. Random projection ( $O(d^{2}w)$ , red dashed) exceeds the target above 50 channels.

Fault	Type	PCA-T	Var	Thr	Full
IDV(1)	A/C feed step	.845	.858	.907	.977
IDV(2)	B comp. step	.814	.867	.908	.970
IDV(4)	Reactor CW step	.723	.739	.736	.980
IDV(5)	Condenser CW step	.774	.722	.797	.975
IDV(6)	A feed loss	.902	.896	.955	.980
IDV(7)	C header press.	.919	.792	.862	.980

TABLE XI: Per-fault F1 at 50% bandwidth on TEP. Bold = best unsupervised. PCA-Triage wins on IDV(7) and IDV(14)—faults where inter-channel correlation shifts are most informative.

$\rho$	PCA-T	Variance	$\Delta$	Uniform
0.00	.795	.852	$-$ .057	.852
0.20	.764	.847	$-$ .084	.839
0.40	.765	.840	$-$ .076	.834
0.60	.762	.837	$-$ .075	.829
0.80	.759	.832	$-$ .073	.828
0.95	.761	.830	$-$ .069	.821

TABLE XII: Synthetic correlation validation. PCA-Triage’s gap vs. Variance narrows as

\rho

increases (from

-0.084

\rho{=}0.2

-0.069

\rho{=}0.95

), consistent with Theorem 1 ’s prediction, but PCA-Triage does not overtake Variance on this simple task.

Config	F1 Mean	F1 Std
Fixed $k{=}3$	.969	.002
Fixed $k{=}5$	.968	.003
Fixed $k{=}10$	.967	.001
Fixed $k{=}15$	.964	.002
Fixed $k{=}20$	.952	.003
Adaptive $k$ (90% var.)	.964	.002

TABLE XIII: Adaptive

k

selection on TEP (50% BW, 3 seeds). Adaptive

k

(90%) matches fixed

k{=}15

, avoiding manual tuning.

Config	F1 Mean	F1 Std
Single $k{=}5$	.968	.003
Single $k{=}10$	.967	.001
Ensemble $[3,5,10]$	.968	.002
Ensemble $[5,10,15]$	.967	.002
Ensemble $[3,5,10,15]$	.968	.002
Hybrid $\alpha{=}0.8$ + sharpening	.970	.001

TABLE XIV: Ensemble scoring on TEP (50% BW, 3 seeds). Ensemble reduces variance but the hybrid PCA+Var scorer remains best.

Dataset	Best Scorer	F1	vs Hybrid PCA+Var
TEP	Hybrid PCA+Var	.970	—
SMD	Hybrid PCA+Var	.987	—
MSL	All tied	.919	$\pm$ 0
PSM	Hybrid PCA+Var	.965	—
HAI	Fixed $k{=}5$	1.000	$+$ 0.0004
SKAB	Fixed $k{=}5$	.562	$+$ 0.013

TABLE XV: Best PCA-Triage scorer per dataset (50% BW, 3 seeds). Hybrid PCA+Var excels on high-correlation datasets; fixed

k{=}5

excels on low-correlation or few-channel datasets.

Dataset	PCA-T	LSTM-Attn	Tfm-Attn	Params
TEP	$\mathbf{.811}$	.623	.631	14K / 12K
SMD	$\mathbf{.980}$	.961	.964	12K / 12K
MSL	$\mathbf{.921}$	.914	.914	15K / 12K
PSM	$\mathbf{.929}$	.886	.884	9K / 12K
HAI	.998	.998	$\mathbf{.999}$	20K / 12K
SKAB	$\mathbf{.541}$	.503	.502	6K / 12K

TABLE XVI: PCA-Triage vs deep learning baselines (50% BW, 20K subsample, 2 seeds). PCA-Triage wins on 5/7 datasets under subsampled conditions with zero trainable parameters and 2–7

\times

faster inference.

Condition	TEP	SMD	PSM
Clean (baseline)	.972	.987	.968
Jitter $\pm 5$	.901	.966	.932
Packet loss 5%	.972	.987	.968
Packet loss 10%	.972	.987	.968
Noise $\sigma{=}0.1$	.966	.986	.968
Noise $\sigma{=}0.3$	.889	.967	.925

TABLE XVII: Real-time robustness: F1 under deployment perturbations (3 datasets

\times

8 conditions, using tuned PCA-Triage configuration). Relative degradation from clean baseline is the key metric.

Method	TEP	SMD	PSM
PCA-Triage (spatial)	.972	.987	.968
Send-on-Delta ( $\delta{=}0.1$ )	.962	.993	.997
Joint PCA+SoD ( $\delta{=}0.1$ )	.964	.990	.974
OGD (regret-optimal)	.934	.976	.921
Variance	.950	.978	.910
Uniform	.922	—	.898

TABLE XVIII: Spatial, temporal, and joint optimization (50% BW, 3 seeds, tuned PCA-Triage configuration). PCA-Triage beats the regret-optimal OGD baseline on all datasets.

Method	Mean Rank
PCA-Triage	1.50
Threshold	2.83
Variance	3.00
Random Dropout	3.50
Uniform	4.17

TABLE XIX: Friedman ranking across 6 datasets at 50% bandwidth (

\chi^{2}=9.33

p=0.053

, Kendall’s

W=0.389

). Lower rank = better. Mutual Info excluded (supervised).

	Threshold	Variance	Uniform	Rand. Drop
$p$ (raw)	.016	.031	.047	.078
Holm sig.	n.s.	n.s.	n.s.	n.s.
W / L / T	6/0/0	5/1/0	5/1/0	5/1/0
Effect $r$	1.00	0.91	0.81	0.71

TABLE XX: Wilcoxon one-sided tests (PCA-Triage

>

baseline, 6 datasets, Holm-corrected). Effect sizes are consistently large despite limited statistical power.

Fig. 13: SKAB sensor traces with anomaly regions (pink). With only 8 channels and limited correlation, all triage methods perform similarly—PCA-Triage’s correlation-exploitation advantage is minimal.

Budget	F1	$\Delta$ F1	BW Saved	MB/hr Saved
10%	0.908	$-$ 5.3%	90%	0.64
30%	0.924	$-$ 3.7%	70%	0.50
50%	0.961	0.0%	50%	0.36
70%	0.963	$+$ 0.2%	30%	0.21
90%	0.963	$+$ 0.2%	10%	0.07

TABLE XXI: Cost-benefit analysis on TEP: F1 degradation vs. bandwidth savings. PCA-Triage maintains F1

>

0.95 down to 30% budget (70% savings).

Adaptive Sampling

Adaptive sampling allocates limited data, measurements, or computation where they are expected to be most useful.

See all 202 papers

LUMIN: Lifecycle-Utility-aware and Metadata-Informed Network for Visual Anomaly Detection
Sep 4, 2026Pengfei YangIndustrial Anomaly Detection Anomaly Detection
Intelligent Precision Instrument, Shenzhen, Guangdong, China
Channel-Adaptive Edge AI: Maximizing Inference Throughput by Adapting Computational Complexity to Channel States
Mar 3, 2026Jierui Zhang, Jianhao Huang, Kaibin HuangAdaptive Inference Edge Inference
Department of Electrical and Computer Engineering, The University of Hong Kong, Hong Kong
(POSTER) From Sensors to Insight: Rapid, Edge-to-Core Application Development for Sensor-Driven Applications
May 4, 2026Komal Thareja, Anirban Mandal, Ewa DeelmanScientific Workflow Automation Edge-Cloud Computing
Renaissance Computing Institute, University of North Carolina Chapel Hill, NC, USA · Information Sciences Institute, University of Southern California, Marina del Rey, CA, USA

Sep 4, 2026J/K move · Enter open · S save

LUMIN: Lifecycle-Utility-aware and Metadata-Informed Network for Visual Anomaly Detection

Pengfei Yang

Intelligent Precision Instrument, Shenzhen, Guangdong, China

Industrial Anomaly Detection Anomaly Detection Efficient Inference

Open paper AskPDF arXiv

Industrial anomaly detection faces two engineering bottlenecks: memory bank construction latency and inference efficiency. Traditional sampling algorithms (Farthest Point Sampling, K-Means, etc.) rely on numerous backbone forward passes and iterative distance computations, with construction times ranging from minutes to hours; heavy computation components such as multi-scale feature extraction struggle to meet the millisecond-level real-time requirements of production lines. This paper focuses on sampling efficiency and inference optimization for industrial deployment with two core contributions: (1) PSP (Plugin Sampler Pipeline)---a four-stage adaptive memory bank sampling pipeline based on 18-dimensional pixel metadata and five complementary visual plugins. PSP completes all sampling with zero backbone forward passes; coarse filtering is sub-second numerical sorting, and metadata extraction is a one-time offline cost. PSP supports progressive deployment and incremental updates. (2) Two engineering optimization strategies---parallel memory bank similarity computation (reducing inference memory and latency by over 95%) and stratified pixel sampling for large-scale evaluation (reducing computation time by 20

\times

while keeping metrics stable). As a vehicle for validation, we introduce LUMIN (Lightweight Universal Manufacturing Inspection Network) with extreme segmentation-head compression, systematically exploring the accuracy-efficiency frontier against strong baselines. Experiments on five benchmarks demonstrate that PSP matches state-of-the-art sampling accuracy at near-random construction cost (341

\times

faster than FPS), while inference optimizations reduce evaluation time by 20

\times

with negligible accuracy loss.

Abstract

Adaptive Sampling

Explore similar work

LUMIN: Lifecycle-Utility-aware and Metadata-Informed Network for Visual Anomaly Detection

Channel-Adaptive Edge AI: Maximizing Inference Throughput by Adapting Computational Complexity to Channel States

(POSTER) From Sensors to Insight: Rapid, Edge-to-Core Application Development for Sensor-Driven Applications