Acoustic sensing for friction and lubrication monitoring involves the use of acoustic emission, guided waves, and ultrasonic reflection or transmission techniques to probe frictional and lubricated interfaces in a non-intrusive manner. It is particularly important for sealed or in-service mechanical systems, where direct access to the contact zone is limited or impossible. However, traditional acoustic sensing approaches often struggle with the complexity, nonlinearity, and non-stationary nature of friction-induced signals, making reliable interpretation and quantitative analysis challenging. Recent advances in artificial intelligence have significantly enhanced the ability to extract meaningful information from complex acoustic signals generated by friction, lubrication, and wear processes. This review surveys the current state of AI-enabled acoustic sensing for friction and lubrication monitoring in tribological systems. The sensing principles and data acquisition characteristics of acoustic emission, guided waves, and ultrasonic techniques are first summarized. Common signal processing and feature representation strategies are then reviewed, followed by key learning-based methods, including regression and classification frameworks, for quantitative prediction of friction, lubricant film thickness, and wear indicators, as well as identification of lubrication regimes, frictional states, and their transitions. Representative engineering applications in automotive systems, energy and power generation equipment, heavy-duty machinery, and sealed or inaccessible frictional interfaces are also discussed, demonstrating the broad applicability of AI-enabled acoustic sensing under practical operating conditions. Recent findings further indicate a growing shift from purely data-driven models toward physics-guided and mechanism-informed learning, with improved robustness, generalization, and interpretability. The review highlights both the capabilities and limitations of existing approaches and outlines emerging opportunities for intelligent and reliable tribological monitoring.

Acoustic emission (AE) signals were recorded during tribological tests on 52,100 steel samples sliding under boundary lubrication conditions in the presence of various chemical additives (an anti-wear agent, two friction modifiers, an extreme pressure additive, and a dispersant). Various parameters of the AE signal were analyzed and compared with the coefficient of friction (CoF), electrical contact resistance (ECR), surface roughness, and wear volume. As expected, the measured CoF, ECR, surface roughness, and wear volume varied between tests due to the action of the additive. Within each test, the variation over time in certain AE frequencies correlated strongly with changes in CoF, while other frequency components correlated with the ECR signal. AE and CoF, surface roughness, and wear volume parameters were also averaged over each test, and the results were plotted against each other while the lubricant additive was varied. These findings revealed that the average certain AE frequencies correlated well with the average CoF, that certain AE frequencies correlated with the surface roughness, and that other average frequency components correlated with the wear volume per test. Together, these results demonstrate that the AE signal contains rich tribological information and is sensitive to asperity interactions and surface film composition. Therefore, AE is a powerful tool for monitoring tribological behavior and lubricant conditions.

It is increasingly important to monitor sliding interfaces within machines, since this is where both energy is lost, and failures occur. Acoustic emission (AE) techniques offer a way to monitor contacts remotely without requiring transparent or electrically conductive materials. However, acoustic data from sliding contacts is notoriously complex and difficult to interpret. Herein, we simultaneously measure coefficient of friction (with a conventional force transducer) and acoustic emission (with a piezoelectric sensor and high acquisition rate digitizer) produced by a steel‒steel rubbing contact. Acquired data is then used to train machine learning (ML) algorithms (e.g., Gaussian process regression (GPR) and support vector machine (SVM)) to correlated acoustic emission with friction. ML training requires the dense AE data to first be reduced in size and a range of processing techniques are assessed for this (e.g., down-sampling, averaging, fast Fourier transforms (FFTs), histograms). Next, fresh, unseen AE data is given to the trained model and the resulting friction predictions are compared with the directly measured friction. There is excellent agreement between the measured and predicted friction when the GPR model is used on AE histogram data, with root mean square (RMS) errors as low as 0.03 and Pearson correlation coefficients reaching 0.8. Moreover, predictions remain accurate despite changes in test conditions such as normal load, reciprocating frequency, and stroke length. This paves the way for remote, acoustic measurements of friction in inaccessible locations within machinery to increase mechanical efficiency and avoid costly failure/needless maintenance.