Trust as a design principle in human–robot collaboration: a review of explainable and adaptive control

What about the human in human robot collaboration?

Human-robot collaborative (HRC) teleoperation requires seamless integration of intention understanding and adaptive control to achieve natural, efficient, and reliable remote manipulation. Existing tele-operation models (TOMs) suffer from limited intention-prediction capabilities, static control parameters, and inadequate adaptation to dynamic operational conditions, resulting in reduced task performance and increased cognitive burden for the operator. The proposed TOM that combines real-time motion intention estimation using long short-term memory (LSTM) + adaptive fuzzy logic control to enhance human-robot collaboration. The proposed TOM leverages multimodal bio-signals, including electromyography, inertial measurement units, and joint kinematics, to decode operator intentions via temporal feature extraction and sequential classification. The LSTM-based classifier processes normalised feature vectors to predict discrete motion intentions with 91.4% accuracy across varying task complexities. Experimental validation using a 6-degree-of-freedom collaborative manipulator and 12 human participants demonstrates significant performance improvements over traditional TOM. The integrated system achieved a 93.5% task-completion success rate, 89% faster execution times, a 60% improvement in placement accuracy, and a 47% reduction in operator mental workload across low, moderate, and high-complexity manipulation tasks. Statistical analysis confirms highly significant improvements (P < 0.001) with large effect sizes across all performance metrics. The proposed model addresses fundamental limitations in HRC teleoperation by providing temporally-aware intention recognition and context-sensitive adaptive control, enabling more natural and efficient collaborative manipulation in remote and hazardous environments.

The widespread implementation of Human–Robot Collaborations (HRC) in industry is becoming an obvious step in the era of Industry 4.0 or even 5.0 and the requirements of sustainable production (SP) due to the increase in productivity and quality while reducing production costs. However, due to limited studies on the establishment of collision thresholds in HRC, in-depth research is needed to determine possible disruptions for safety in HRC. So, firstly, in order to identify current issues and critical challenges in the field of HRC safety, an analysis of the literature was carried out using the Systematic Literature Review (SLR) & Mapping in Literature Reviews (MLR) methodology based on the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines (PRISMA). A total of 4,107 research articles were identified in Science Direct—Elsevier, Web of Science and Springer, that focussed on safety in HRC in SP. In this study, 103 items were then analysed and 53 in the second stage. The key research contributions are developed as the principles for enhancing safety in HRC in industry: (1) improving HRC workspace safety, (2) building mechanisms for monitoring HRC workspaces and (3) managing ethical issues. In line with these trends, secondly, experimental studies were provided. Next, empirical research was conducted to establish the parameters for ensuring the safety of the monitoring system for work in HRC. Therefore, analysis of the video recordings on the efficiency of detecting collision in HRC, when using the Region-Based Convolutional Neural Network (YOLOv8 Tiny), within the vision system were researched. The experimental studies indicate the necessary actions within a monitoring system in HRC in order to ensure workplace safety in the HRC environment. Finally, the research results suggest future research directions as (1) proactive HRC with attention to collaborative robots and (2) integration with the approach, to ensure safety in HRC into SP strategy in industry. This study provides new insights into HRC and gives an outlook for development of safety issues in HRC in industry.

This paper presents a comprehensive framework for evaluating the robustness and adaptability of human–robot collaboration (HRC) controllers under a spectrum of dynamic and unpredictable human intentions. Building upon variable admittance controller (VAC) frameworks augmented with Radial Basis Function Neural Network (RBFNN) online adaptation, we introduce two key innovations: (1) an intent-aware human force generator capable of simulating aggressive, hesitant, oscillatory, conflicting, and nominal behaviors, through the modulation of force gains and the introduction of stochastic noise, and (2) the extension of VAC to incorporate variable stiffness as an adaptive control parameter alongside damping and inertia. The adaptive parameters are jointly tuned online using a self-supervised learning (SSL) mechanism driven by motion error metrics and interaction dynamics. The framework is simulated in a dual-arm collaborative manipulation scenario involving two 7-DoF Franka Emika Panda robots transporting a shared object in a high-fidelity simulation environment. Simulation results demonstrate the system’s capability to maintain stable behavior and minimize tracking error despite abrupt changes in human intent. This work provides a novel and systematic tool for stress-testing adaptive controllers in HRC, with implications for the design of resilient, safe, and reliable robotic systems in real-world collaborative environments.

Compliant adaptive control of human upper-limb exoskeleton robot with unknown dynamics based on a Modified Function Approximation Technique (MFAT)

In this paper, we develop a predictive safety control framework for robotic exoskeletons that integrates Kalman smoothing (KS) for noise reduction in sensor data, long short-term memory (LSTM)-based force prediction, adaptive impedance control, and dynamic barrier functions. The proposed approach is applicable to safety-critical tasks in uncertain environments and addresses important limitations of prior methods including disjointed perception control pipelines, rigid safety constraints, and lack of guaranteed stability. The main contributions of the proposed approach are: a unique KS-LSTM coupling to reduce interaction force prediction errors by 44.5% while operating in real-time (22.3 ms/cycle); adaptive control that modulates stiffness/damping parameters using LSTM-predicted forces; and context-aware barrier functions with formal globally uniformly ultimately bounded (GUUB) stability guarantees via Lyapunov analysis. Through extensive simulations, the proposed approach demonstrates superior performance over baselines in prediction accuracy, trajectory tracking, and constraint compliance (95.04%), especially in obstacle avoidance and human–robot collaboration. This framework connects learning-based prediction and real-time certifiable control, which helps to ensure safer physical human–robot interaction in rehabilitation, industrial, and assistive settings.

The warehouse picking process is one of the most critical components of logistics operations. Human–robot collaboration (HRC) is seen as an important trend in warehouse picking, as it combines the strengths of both humans and robots in the picking process. However, in current human–robot collaboration frameworks, there is a lack of effective communication between humans and robots, which results in inefficient task execution during the picking process. To address this, this paper considers trust as a communication bridge between humans and robots and proposes the Stackelberg trust-based human–robot collaboration framework for warehouse picking, aiming to achieve efficient and effective human–robot collaborative picking. In this framework, HRC with trust for warehouse picking is defined as the Partially Observable Stochastic Game (POSG) model. We model human fatigue with the logistic function and incorporate its impact on the efficiency reward function of the POSG. Based on the POSG model, belief space is used to assess human trust, and human strategies are formed. An iterative Stackelberg trust strategy generation (ISTSG) algorithm is designed to achieve the optimal long-term collaboration benefits between humans and robots, which is solved by the Bellman equation. The generated human–robot decision profile is formalized as a Partially Observable Markov Decision Process (POMDP), and the properties of human–robot collaboration are specified as PCTL (probabilistic computation tree logic) with rewards, such as efficiency, accuracy, trust, and human fatigue. The probabilistic model checker PRISM is exploited to verify and analyze the corresponding properties of the POMDP. We take the popular human–robot collaboration robot TORU as a case study. The experimental results show that our framework improves the efficiency of human–robot collaboration for warehouse picking and reduces worker fatigue while ensuring the required accuracy of human–robot collaboration.

With the rapid advancement of science and technology, robotics is evolving towards more profound and extensive applications. Nevertheless, the inherent limitations of traditional industrial “caged” robots have significantly impeded the full utilization of their capabilities. Consequently, breaking free from these constraints and realizing human–robot collaboration has emerged as a new developmental trend in the robotics field. The collision-response mechanism, as a crucial safeguard for human–robot collaboration safety, has become a pivotal issue in enhancing the performance of human–robot interaction. To address this, an adaptive admittance control collision-response algorithm is proposed in this paper, grounded in the principle of admittance control. A collision simulation model of the AUBO-i5 collaborative robot is constructed. The effectiveness of the proposed algorithm is verified through simulation experiments focusing on both the end-effector collision and body collision of the robot, and by comparing it with existing admittance control algorithms. Furthermore, a collision-response experimental platform is established based on the AUBO-i5 collaborative robot. Experimental studies on end-effector and body collisions are conducted, providing practical validation of the reliability and utility of the proposed adaptive admittance control collision-response algorithm.

PurposeIn customized production such as plate workpiece grinding, because of the diversity of the workpiece shapes and the positional/orientational clamping errors, great efforts are taken to repeatedly calibrate and program the robots. To change this situation, the purpose of this study is to propose a method of robotic direct grinding for unknown workpiece contour based on adaptive constant force control and human–robot collaboration.Design/methodology/approachFirst, an adaptive constant force controller based on stiffness estimation is proposed, which can distinguish the contact of the human hand and the unknown workpiece contour. Second, a normal vector search algorithm is developed to calculate the normal vector of the unknown workpiece contour in real-time. Finally, the force and position are controlled in the calculated normal and tangential directions to realize the direct grinding.FindingsThe method considers the disturbance of the tangential grinding force and the friction, so the robot can track and grind the workpiece contour simultaneously. The experiments prove that the method can ensure the force error and the normal vector calculating error within 0.3 N and 4°. This human–robot collaboration pattern improves the convenience of the grinding process.Research limitations/implicationsThe proposed method realizes constant force grinding of unknown workpiece contour in real-time and ensures the grinding consistency. In addition, combined with human–robot collaboration, the method saves the time spent in repeated calibration and programming.Originality/valueCompared with other related research, this method has better accuracy and anti-disturbance capability of force control and normal vector calculation during the actual grinding process.

We developed a human–robot collaborative manipulation system (co-manipulation system) in the form of a power assist robotic system (PARS) where a human and a robot collaborated to perform the co-manipulation of an object with power assistance. We conducted an experiment (the first experiment), where in each trial of the experiment, a human subject performed the co-manipulation of the object with the PARS, and an expert human–robot co-manipulation researcher observed the co-manipulation task. We collected the co-manipulation and observation data, analyzed the data, and conducted reviews of the related literature, and developed the HRC (human–robot collaboration) metrology, which consisted of necessary criteria, metrics and methods to assess human–robot collaborative manipulation tasks. The proposed HRC metrology consisted of both human–robot collaborative performance and human–robot interactions (HRI) related assessment criteria. Then, we developed another human–robot co-manipulation system using a robot manipulator. In this system, the human–robot co-manipulation task was performed in conjunction with a collaborative assembly task between the robot and human co-workers. In another experiment (the second experiment), we assessed the co-manipulation task for each robotic system separately based on the developed HRC metrology (set of assessment criteria, metrics and methods) to verify and validate the practicality, usability and effectiveness of the criteria, metrics and methods. The results showed that the HRC metrology was effective and practical in assessing the co-manipulation tasks. We then discussed the strengths and limitations of the assessment criteria, metrics and methods. The proposed HRC metrology can be used to assess human–robot collaborative performance and human–robot interactions in human–robot co-manipulation tasks with potential real-world applications in industrial manipulation and manufacturing, transport, logistics, civil construction, rescue and disaster management, timber processing, etc.

Recent developments in robotics allow the design of work systems with enhanced human robot collaboration (HRC) for assembly tasks. Productivity improvements are a common aim for companies that look into the implementation of HRC. To harvest the full productivity potential of these work systems, an analysis of the HRC work processes is essential. However, a dedicated method for the analysis of productivity in HRC is missing. This paper presents an approach using 3D-cameras to observe the employee in HRC. The approach links this information to robot states. The resulting analysis aims at improving the productivity of the work system e.g. by identifying and reducing balancing losses in HRC. The method tracks the movements of the employees in the HRC area and matches the corresponding robot activities.

Industrial robots are undergoing a transformative shift as Artificial Intelligence (AI)-driven and bio-inspired control strategies unlock new levels of precision, adaptability, and multi-dimensional sustainability aligned with Industry 5.0 (energy efficiency, material circularity, and life-cycle emissions). This systematic review analyzes 160 peer-reviewed industrial robotics control studies (2023–2025), including an expanded bio-inspired/human-centric subset, to evaluate: (1) the dominant and emerging control methodologies; (2) the transformative role of digital twins and 5G-enabled connectivity; and (3) the persistent technical, ethical, and environmental challenges. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, the study employs a rigorous methodology, focusing on adaptive control, deep reinforcement learning (DRL), human–robot collaboration (HRC), and quantum-inspired algorithms. The key findings highlight up to 30% latency reductions in real-time optimization, up to 22% efficiency gains through digital twins, and up to 25% energy savings from bio-inspired designs (all percentage ranges are reported relative to the comparator baselines specified in the cited sources). However, critical barriers remain, including scalability limitations (with up to 40% higher computational demands) and cybersecurity vulnerabilities (with up to 20% exposure rates). The convergence of AI, bio-inspired systems, and quantum computing is poised to enable sustainable, autonomous, and human-centric robotics, yet requires standardized safety frameworks and hybrid architectures to fully support the transition from Industry 4.0 to Industry 5.0. This review offers a strategic roadmap for future research and industrial adoption, emphasizing human-centric design, ethical frameworks, and circular-economy principles to address global manufacturing challenges.

Human-Robot Collaboration (HRC) refers to the interaction of workers and robots in a shared workspace. Owing to the integration of the industrial automation strengths with the inimitable cognitive capabilities of humans, HRC is paramount to move towards advanced and sustainable production systems. Although the overall safety of collaborative robotics has increased over time, further research efforts are needed to allow humans to operate alongside robots, with awareness and trust. Numerous safety concerns are open, and either new or enhanced technical, procedural and organizational measures have to be investigated to design and implement inherently safe and ergonomic automation solutions, aligning the systems performance and the human safety. Therefore, a bibliometric analysis and a literature review are carried out in the present paper to provide a comprehensive overview of Occupational Health and Safety (OHS) issues in HRC. As a result, the most researched topics and application areas, and the possible future lines of research are identified. Reviewed articles stress the central role played by humans during collaboration, underlining the need to integrate the human factor in the hazard analysis and risk assessment. Human-centered design and cognitive engineering principles also require further investigations to increase the worker acceptance and trust during collaboration. Deepened studies are compulsory in the healthcare sector, to investigate the social and ethical implications of HRC. Whatever the application context is, the implementation of more and more advanced technologies is fundamental to overcome the current HRC safety concerns, designing low-risk HRC systems while ensuring the system productivity.

SUMMARYThis work aims to propose an innovative mechanism of human–robot collaboration (HRC) for mobile service robots in the application of elderly and disabled assistance. Previous studies on HRC mechanism usually focused on integrating decision-making intelligence of human beings by qualitative judgment and reasoning intelligence of robots by quantitative calculation. Instead, novelties of the proposed methodology include (1) constructing an HRC framework by taking reference from the Adaptive Control of Thought – Rational (ACT-R) human cognitive architecture; (2) establishing semantic webs of cognitive reasoning through human–robot interaction (HRI) and HRC to plan and implement complex tasks; and (3) realizing human–robot intelligence fusion by mutual encouragement, connect, and integration of modules of human, robot, perception, HRI, and HRC in the ACT-R architecture. Its technical feasibility is validated by some selected experiments within a “pouring” scenario. Further, although this study is oriented to mobile service robots, the modularized design of hardware and software makes its extensive use feasible in other types of service robots like smart rehabilitation beds, wheelchairs, and cleaning equipments.

In this study, a smart and affordable system that utilizes an RGB-D camera to measure the exact position of an operator with respect to an adjacent robotic manipulator was developed. This developed technology was implemented in a simulated human operation in an automated manufacturing robot to achieve two goals; enhancing the safety measures around the robot by adding an affordable smart system for human detection and robot control and developing a system that will allow the between the human-robot collaboration to finish a predefined task. The system utilized an Xbox Kinect V2 sensor/camera and Scorbot ER-V Plus to model and mimics the selected applications. To achieve these goals, a geometric model for the Scorbot and Xbox Kinect V2 was developed, a robotics joint calibration was applied, an algorithm of background segmentation was utilized to detect the operator and a dynamic binary mask for the robot was implemented, and the efficiency of both systems based on the response time and localization error was analyzed. The first application of the Add-on Safety Device aims to monitor the working-space and control the robot to avoid any collisions when an operator enters or gets closer. This application will reduced and remove physical barriers around the robots, expand the physical work area, reduce the proximity limitations, and enhance the human-robots interaction (HRI) in an industrial environment while sustaining a low cost. The system was able to respond to human intrusion to prevent any collision within 500 ms on average, and it was found that the system’s bottleneck was PC and robot inter-communication speed. The second application was developing a successful collaborative scenario between a robot and a human operator, where a robot will deposit an object on the operator’s hand, mimicking a real-life human-robot collaboration (HRC) tasks. The system was able to detect the operator’s hand and it’s location then command the robot to place an object on the hand, the system was able to place the object within a mean error of 2.4 cm, and the limitation of this system was the internal variables and data transmitting speed between the robot controller and main computer. These results are encouraging and ongoing work aims to experiment with different operations and implement gesture detection in real-time collaboration tasks while keeping the human operator safe and predicting their behavior.