The integration of Artificial Intelligence (AI) into manufacturing is transforming the industry by advancing predictive maintenance, quality control, and supply chain optimisation, while also driving the shift from Industry 4.0 towards a more human-centric and sustainable vision. This emerging paradigm, known as Industry 5.0, emphasises resilience, ethical innovation, and the symbiosis between humans and intelligent systems, with AI playing a central enabling role. However, challenges such as the “black box” nature of AI models, data biases, ethical concerns, and the lack of robust frameworks for trustworthiness hinder its widespread adoption. This paper provides a comprehensive survey of AI trustworthiness in the manufacturing industry, examining the evolution of industrial paradigms, identifying key barriers to AI adoption, and examining principles such as transparency, fairness, robustness, and accountability. It offers a detailed summary of existing toolkits and methodologies for explainability, bias mitigation, and robustness, which are essential for fostering trust in AI systems. Additionally, this paper examines challenges throughout the AI pipeline, from data collection to model deployment, and concludes with recommendations and research questions aimed at addressing these issues. By offering actionable insights, this study aims to guide researchers, practitioners, and policymakers in developing ethical and reliable AI systems that align with the principles of Industry 5.0, ensuring both technological advancement and societal value.

Affective computing involves human–computer interaction (HCI) where an interface can detect and respond in context to a user’s emotions. Emotions play an essential role in the daily activities of a human during work and, as a consequence, can critically affect their decision-making. Recent research on affective computing opens the door to exploring the emotional aspects of computer aided tasks and HCI; however, there is little research in this area related to computer aided design (CAD). There is an established connection between emotional and cognitive processes and without these emotional markers decision-making in humans is almost non-existent.The holy grail of engineering design is for a concept to be ‘right-first-time’ and only by delving into the affect-cognitive domain will insights in terms of the associability of critical decision-making and the underlying intent of engineering decisions and tasks be better understood. Work in this field aims to recognize, interpret and process human emotions in order to support and adapt computer-based systems and their associated interfaces to user behaviors and to elicit appropriate system responses. This is a very contemporary area in engineering design research. One key aspect of provisioning CAD systems with the capability to detect the emotional state of the engineer is to lessen the chance of distorted reasoning as a result of irrational judgements. Work in this area can potentially identify emotional markers that influence the coherence/incoherence of reasoning against knowledge as well as the potential quality and suitability of a specific engineering solution or approach.This paper proposes and investigates a methodology to determine the emotional aspects attributed to a set of CAD design tasks by analyzing the CAD operators’ psycho-physiological signals. Two sets of experiments were conducted. A pilot case study focused on modeling tasks carried out in Solid Edge™ was run to prove the validity of the methodology, the tools applied and subsequent analysis. The main case study extended this to more wide-ranging user logging, incorporating a series of configuration and optimization tasks in Siemens NX™ for a real design task. Psycho-physiological signals of electroencephalography (EEG) and galvanic skin resistance (GSR)/electrocardiography (ECG) were recorded along with a log of CAD system user interactions. A fuzzy logic model was established to map the psycho-physiological signals to a set of key emotions, namely frustration, satisfaction, engagement and challenge and the results analyzed.The methodology and subsequent analysis applied was found to be repeatable. Over the various task action-chains the elicitation and interpretation of each participant’s emotions were successfully carried out with significant correlations demonstrated between the associated engineers’ CAD activities and their reported emotional states

Deep neural network models have become powerful tools of machine learning and artificial intelligence. They can approximate functions and dynamics by learning from examples. This paper reviews the state-of-art of deep learning-based object detection frameworks that are used for fruit detection in general and for olive fruit in particular. A dataset of olive fruit on the tree is built to train and evaluate deep models. The ultimate goal of this work is the capability of on-edge real-time olive fruit detection on the tree from digital videos. Recent work in deep neural networks has led to the development of a state-of-the-art object detector termed You Only Look Once version five (YOLOv5). This paper builds a dataset of 1.2 K source images of olive fruit on the tree and evaluates the latest object detection algorithms focusing on variants of YOLOv5 and YOLOR. The results of the YOLOv5 models show that the YOLOv5 new network models are able to extract rich olive features from images and detect the olive fruit with a high precision of higher than 0.75 mAP_0.5. YOLOv5s performs better for real-time olive fruit detection on the tree over other YOLOv5 variants and YOLOR.

Artificial intelligence (AI) is increasingly adopted in manufacturing for tasks such as automated inspection, predictive maintenance, and condition monitoring. However, the opaque, black-box nature of many AI models remains a major barrier to industrial trust, acceptance, and regulatory compliance. This study investigates how explainable artificial intelligence (XAI) techniques can be used to systematically open and interpret the internal reasoning of AI systems commonly deployed in manufacturing, rather than to optimise or compare model performance. A unified explainability-centred framework is proposed and applied across three representative manufacturing use cases encompassing heterogeneous data modalities and learning paradigms: vision-based classification of casting defects, vision-based localisation of metal surface defects, and unsupervised acoustic anomaly detection for machine condition monitoring. Diverse models are intentionally employed as representative black-box decision-makers to evaluate whether XAI methods can provide consistent, physically meaningful explanations independent of model architecture, task formulation, or supervision strategy. A range of established XAI techniques, including Grad-CAM, Integrated Gradients, Saliency Maps, Occlusion Sensitivity, and SHAP, are applied to expose model attention, feature relevance, and decision drivers across visual and acoustic domains. The results demonstrate that XAI enables alignment between model behaviour and physically interpretable defect and fault mechanisms, supporting transparent, auditable, and human-interpretable decision-making. By positioning explainability as a core operational requirement rather than a post hoc visual aid, this work contributes a cross-modal framework for trustworthy AI in manufacturing, aligned with Industry 5.0 principles, human-in-the-loop oversight, and emerging expectations for transparent and accountable industrial AI systems.

This paper presents sub-6 GHz channel measurements using a directional antenna at the transmitter and a directional or omnidirectional antenna at the receiver at 4.145 GHz in sparse and dense industrial environments for a line-of-sight scenario. Furthermore, the first measured over-the-air error vector magnitude (EVM) results depending on different 5G new radio modulation and coding schemes (MCSs of16 QAM, 64 QAM and 256 QAM) are provided. From the measurement campaigns, the path loss exponents (PLE) using a directional and an omnidirectional antenna at the receiver in the sparse and the dense environment are 1.24/1.39 and 1.35/1.5, respectively. PLE results are lower than the theoretical free space PLE of 2, indicating that indoor industrial environments have rich multipaths. The measured power delay profiles show the maximum root mean square (RMS) delay spreads of 11 ns with a directional antenna and 34 ns with an omnidirectional antenna at the receiver in a sparse industrial environment. However, in a dense industrial environment the maximum RMS delay spreads are significantly increased: maximum RMS delay spreads range from 226 to 282 ns for the omnidirectional and the directional antenna configuration. EVM measurements show that to increase coverage and enable higher MCS modes to be used for reliable data transmission, in both industrial environments using a directional antenna at the transmitter and the receiver is required. The large-scale path loss models, multipath time dispersion characteristics and EVM results provide insight into the deployments of 5G networks operating at sub-6 GHz frequency bands in different industrial environments.