Featured Application The proposed method enables the adaptive correction of robotic welding trajectories based on the 3D scanning of real components, reducing defects and rework in industrial processes characterized by high geometric variability.Abstract The geometric variability of industrial components represents a persistent challenge in robotic arc welding, particularly in high-volume manufacturing environments where parts are positioned in fixtures based on nominal CAD assumptions. Even moderate deviations in dimensions or seating conditions can lead to weld defects, rework, and reduced process capability when conventional offline programming is employed. This paper presents an applied industrial workflow for adaptive robotic welding trajectory correction that integrates full-field 3D optical metrology with a data-driven deep reinforcement learning (DRL) model. Prior to welding, each component is scanned using a structured-light 3D system, and critical geometric deviations are extracted relative to the nominal CAD model. These deviations define a compact state representation that is mapped, via a trained DRL agent, to corrective translational and rotational adjustments of the welding trajectory. Importantly, all trajectory corrections are computed offline, ensuring compatibility with standard industrial robot controllers and avoiding real-time computational overheads. The proposed approach is validated using real production data from an industrial batch of 5000 components characterized by significant dimensional variability and limited process capability. Experimental results demonstrate a reduction in welding defects exceeding 90%, elimination of rework associated with improper part positioning, and an improvement of the overall process performance to a sigma level of 5.219. The results show that combining 3D optical metrology with learning-based trajectory adaptation enables robust compensation of part-level geometric deviations without mechanical fixture modifications. The proposed method provides a practical and scalable solution for improving welding quality in manufacturing environments affected by upstream variability and imperfect part positioning.

In the digital age, artificial intelligence (AI) brings a new personalization and development of logical-creative thinking in student education. This booming technological field brings new opportunities and resources for higher education. In this context, the paper presents the project for a new master's program. The mission and objectives of the master's program Applied Artificial Intelligence in Mechatronics and Robotics, taught in English, are in line with the mission of the higher education institution and with the educational requirements identified on the labour market, and propose innovative directions for the development of educational offers necessary in a dynamic labour market. The design of the learning and teaching process would have been made in such a way as to allow the choice of flexible learning paths through optional subjects and in compliance with the procedures in force. Details related to the curriculum and the expected learning outcomes are presented in the paper. Also, the connection between learning outcomes and the requirements of the professions envisaged for graduates is discussed.

The Mechatronics Education Days (ZEM - Zilele Educatiei Mecatronice) is a prestigious event attended by all the universities in Romania that provide academic training in the field of mechatronics. The event is a pole of attraction and inspiration for the students of universities in the country and beyond. It is the most important event at which students can verify and test their skills by participating in competitions with practical applicability for the theoretical knowledge accumulated during their academic training. This competition is annually organised. In 2025, this competition was organised by the University of Craiova. The organisation of such a competition poses complex technical and organisational problems. This paper presents how these problems were identified and for which appropriate solutions were proposed for the 2025 competition.

The rapid growth of the real estate market has led to the appearance of more and more residential areas and large apartment buildings that need to be managed and maintained by a single real estate developer or company. This scientific article details the development of a novel method for inspecting buildings in a semi-automated manner, thereby reducing the time needed to assess the requirements for the maintenance of a building. This paper focuses on the development of an application which has the purpose of detecting imperfections in a range of building sections using a combination of machine learning techniques and 3D scanning methodologies. This research focuses on the design and development of a machine learning-based application that utilizes the Python programming language and the PyTorch library; it builds on the team′s previous study, in which they investigated the possibility of applying their expertise in creating construction-related applications for real-life situations. Using the Zed camera system, real-life pictures of various building components were used, along with stock images when needed, to train an artificial intelligence model that could identify surface damage or defects such as cracks and differentiate between naturally occurring elements such as shadows or stains. One of the goals is to develop an application that can identify defects in real time while using readily available tools in order to ensure a practical and affordable solution. The findings of this study have the potential to greatly enhance the availability of defect detection procedures in the construction sector, which will result in better building maintenance and structural integrity.

Autonomous legged navigation in unstructured environments is still an open problem which requires the ability of an intelligent agent to detect and react to potential obstacles found in its area. These obstacles may range from vehicles, pedestrians, or immovable objects in a structured environment, like in highway or city navigation, to unpredictable static and dynamic obstacles in the case of navigating in an unstructured environment, such as a forest road. The latter scenario is usually more difficult to handle, due to the higher unpredictability. In this paper, we propose a vision dynamics approach to the path planning and navigation problem for a quadruped robot, which navigates in an unstructured environment, more specifically on a forest road. Our vision dynamics approach is based on a recurrent neural network that uses an RGB-D sensor as its source of data, constructing sequences of previous depth sensor observations and predicting future observations over a finite time span. We compare our approach with other state-of-the-art methods in obstacle-driven path planning algorithms and perform ablation studies to analyze the impact of architectural changes to our model components, demonstrating that our approach achieves superior performance in terms of successfully generating collision-free trajectories for the intelligent agent.

Researchers can now utilize new materials to create innovative models for lower limb prostheses and explore novel ways to use them for efficient dynamic control. To achieve user-friendliness, one area of research focuses on recovering and reusing kinetic walking energy for dynamic control. This paper proposes a new design for a magnetorheological (MR) valve, along with a rotary actuator which offers a dynamic control for a lower limb prosthesis. The design will allow the storage of the energy during heel and mid-foot contact phases and to utilize it during toe support to lift the foot off the ground and establish a balance for the lower limb prosthesis. The energy is transferred through a magnetorheological hydraulic circuit and stored using a pneumatic system. The speed of energy transfer is regulated by magnetorheological valves. A series of MR valve designs were proposed and evaluated experimentally, which allowed the identification of the most suitable variant in the targeted application context. The design of the lower limb prosthesis was simulated using SolidWorks, and its dynamic behavior was analyzed in ANSYS.

The gap between industry and higher education refers to the disconnect or lack of collaboration between academic institutions and the industries that they serve. This gap can manifest in various ways, such as a mismatch between the skills taught in the classroom and the skills needed in the workplace, a lack of industry input in curriculum development, or a lack of opportunities for students to gain practical experience and interact with industry professionals. One of the major factors contributing to the gap is the pace of technological advancement. Industry moves at a faster pace than academia, and it can be challenging for academic institutions to keep up with the latest developments in their fields. This can result in a delay in the integration of new technologies into academic curricula, leading to a skills gap between graduates and industry requirements. In this paper we will present the result of the collaboration between our institution, and two representant automotive companies, for developing a new course "Mechatronics in Automotive".

The Internet of Things (IoT) is the natural response to the need of collecting remote data, to monitor and to remote control hardware devices, given the advancements in the telecommunication field. Started in 1980’s as an application to remotely monitor the supply level of a vending machine placed in a university campus, coined as term by Kevin Aston in 1999, the IoT was used by more than 25% of businesses at 2019 level, and gained popularity as a remote tool for home appliances. A branch of IoT is represented by urban IoT systems, designed to support the smart city vision - Smart City, which seeks to exploit the most advanced communication technologies to support value-added services for city administration and citizens; up until the present, the main developments in this area was conducted toward smart traffic / parking management, and air quality monitoring. In this context, the SmartTest research project aims to create a prototype equipment in form of an intelligent modular system for metrological calibration/verification for time and for distance parameters in urban transport, with robotic assistance and IoT functionalities, which will lead to the development of innovative products and services intended for metrological verifications in motor vehicle traffic – transport of goods but also of passengers.

The impact of two link kinematic chain with a flat surface is studied. The open chain has one impacting link with different incident angles and initial velocities. Computer simulations are conducted, and the results are experimentally verified with a high-speed camera. Pre-impact and post-impact velocities of each link are calculated from the experimental data. The frictional impact of two link kinematic chain is mathematically modeled and experimentally validated. Post-impact kinetic energy allocation of each link and the system is investigated. The friction force influence at the contact point is analyzed. For some impact angles, the kinetic energy of the non-impacting link is increasing.

This paper describes the implementation of a solution for detecting the machining defects from an engine block, in the piston chamber. The solution was developed for an automotive manufacturer and the main goal of the implementation is the replacement of the visual inspection performed by a human operator with a computer vision application. We started by exploring different machine vision applications used in the manufacturing environment for several types of operations, and how machine learning is being used in robotic industrial applications. The solution implementation is re-using hardware that is already available at the manufacturing plant and decommissioned from another system. The re-used components are the cameras, the IO (Input/Output) Ethernet module, sensors, cables, and other accessories. The hardware will be used in the acquisition of the images, and for processing, a new system will be implemented with a human–machine interface, user controls, and communication with the main production line. Main results and conclusions highlight the efficiency of the CCD (charged-coupled device) sensors in the manufacturing environment and the robustness of the machine learning algorithms (convolutional neural networks) implemented in computer vision applications (thresholding and regions of interest).

The Internet of Things (IoT) is the natural response to the need of collecting remote data, to monitor and to remote control hardware devices, given the advancements in the telecommunication field. Started in 1980’s as an application to remotely monitor the supply level of a vending machine placed in a university campus, coined as term by Kevin Aston in 1999, the IoT was used by more than 25% of businesses at 2019 level, and gained popularity as a remote tool for home appliances. A branch of IoT is represented by urban IoT systems, designed to support the smart city vision - Smart City, which seeks to exploit the most advanced communication technologies to support value-added services for city administration and citizens; up until the present, the main developments in this area was conducted toward smart traffic / parking management, and air quality monitoring.

Establishing procedures and equipment for more accurate evaluation of the brake system performances is a major goal for automakers, but also for researchers, test equipment manufacturers, racing drivers and their staff, or for experts and forensics in their activity of accidents reconstruction. Increasing the performance and complexity of new vehicle stability control systems as well as driver assistance systems require finding new solutions for measuring and determining braking and stopping time and space, as well as a lot of new parameters such as braking system response, driver’s reaction times, actual value of the traction grip, corrections due to the slope of the road, or to the differences between the actual speed and the one displayed on board. Changes in adhesion and normal load on each wheel contact patch that occur under certain driving conditions develop lateral forces and yaw that disrupt longitudinal dynamics and must be determined and considered. The lack of braking marks in the case of partially braked or unlocked wheel by the ABS system is also a major impediment to the correct and accurate reconstitution of the braking and stopping space or speed. This paper describes a series of experimental research carried out by the authors using a professional GPS device, IMU, infrared thermal camera and a series of sensors that allow to determine with high precision the trajectory, radius of road curvature and tilt, the vehicle real speed, longitudinal and lateral accelerations, yaw speed or drift angle. Based on these determinations, several braking and acceleration parameters were estimated in alignment, in curves or in ramp and slope, respectively in coast down mode.

The use of image acquisition, processing and recognition techniques already has a history in terms of practical applications. The current and future development of industrial applications favors the development of such techniques based on increasing their performance. Considering this trend, it is necessary to adapt the contents of the artificial vision courses with an emphasis on the practical applicability with the high performance of the new methods and techniques in the field. At the Faculty of Automation, Computers and Electronics, from the University of Craiova, Romania, such a course is offered to students from the undergraduate programs in Multimedia Systems Engineering, Applied Electronics, and Mechatronics and Robotics, respectively. The course includes chapters dedicated to digital image acquisition, image processing, image segmentation, image descriptors, image classification and recognition, and applications. This paper will present how to upgrade the chapter related to applications, referring to the practical application developed by a group of doctoral students from our faculty. Methodologies for the development of reliable and complex computer vision applications which are used in a manufacturing environment are generally presented. The principles for the V-model development methodology and the Agile methodology are parts of this presentation. Students will receive the basic knowledge needed for comparing the advantages and disadvantages of these established methods in the manufacturing industry. In practice, these methods should measure the reliability of the system, what percentage of the functional requirements is achieved and at what quality, the behavior of the hardware and software components, and the behavior of the system when integrated into the plant environment. For validating the concept, the results obtained using the V-model and Agile methodologies into a computer vision automated inspection application for engine blocks from an automotive production plant will be exemplified.

Estimation of the power loss in miniature ball bearing grease lubricated is a complex problem. Usually the applied loads (radial and axial) have small values and the methodologies recommended by the bearing companies are cannot applied for these conditions. For a ball bearing, some friction processes have differential contribution to the total friction torque and power loss. For very low loads, the lubricant is the most important source for friction torque. In the present paper the authors determined experimentally the friction torque both in a standard 7000C angular contact ball bearing (ACBB) and a modified 7000C ACBB containing only 3 balls without cage, operating with very low axial load and lubricated with lithium soap grease. The experimental values of the friction torque have been correlated with the theoretical Houpert’s IVR model developed for hydrodynamic rolling resistances in ball race contacts considering the viscosity of the base oil of grease.

For people with amputated limbs, it is necessary to make high-performance prostheses that reproduce, as accurately as possible, the functions of the amputated limb. In order to achieve them, a preliminary study of the limbs from a kinematic and dynamic point of view is necessary. In this paper, an acquisition system for the kinematic and dynamic parameters of the legs is proposed. It consists of a sensory system attached to the legs and an acquisition unit built around a microcontroller. The sensory system has two subcomponents. A sensory system for determining the distribution of body weight on the sole, made of resistive pressure sensors. A second sensory system determines the kinematics and dynamics of the legs while walking, based on a data fusion between gyroscopic and accelerometer sensors. The acquired data is transmitted in real-time, via wi-fi, to a computer system for interpretation. After processing and interpreting the data using standard data sets for comparison, the position of the legs, the type of gait, and the phase of movement can be determined. Constructive, the system is configurable and can be adapted to any person, male or female, regardless of shoe size.

For people with amputated lower limbs, it is imperative to make high-performance prostheses that reproduce, as accurately as possible, the functions of the amputated limb. In this case, a preliminary study of the lower limbs from a kinematic and dynamic point of view is necessary. This paper proposes a prosthesis design and a system for acquiring the information needed to determine the stepping phase kinematic and dynamic parameters of the legs. This system consists of a sensory system attached to the legs and a acquisition data unit built around a microcontroller. The sensory system is based on a sensory system for determining the weight distribution on the sole, made of resistive pressure sensors. The sensory system will be subjected to measurement repeatability and homogeneity tests to evaluate and validate the accuracy and error of the proposed solution. The data obtained by the sensory system is transmitted in real-time, via wi-fi, to a computer system for interpretation. After processing and interpreting the data using standard data sets for comparison, the position of the legs, the type of gait and the phase of movement can be determined. Constructively, the system is configurable and can be adapted to any person, male or female, regardless of shoe size.

The impact of artificial intelligence applications in today’s society is constantly growing. In this context, it is important to adapt the curricula of artificial intelligence courses with an emphasis on the practical applicability of new concepts and methods. At the University of Craiova, Romania, such a course is offered to undergraduate students in Robotics and Automation. The course includes chapters dedicated to problem solving, fuzzy logic, neural networks, and expert systems, with the identification of techniques related to the representation of knowledge and deep learning, respectively. This paper will present how to update the chapter on deep learning, referring to practical applications developed by groups of teachers, researchers and students from the faculty. The following applications are considered. An application is related to the management of a mobile robot that helps people with special needs in terms of control of the lower and upper limbs but who also have specific problems related to their visual system. Another application refers to the use of artificial intelligence techniques in applications for detecting casting defects in car engines. Another chapter that is to be introduced in the course refers to bots - cooperative software robots. The introduction of this chapter is also related to the international presence of an important number of highly successful companies that have been established by Romanian entrepreneurs. The three applications are demonstrated during the practical classes of the artificial intelligence course. Latest hardware platforms (i.e. nVidia Jetson Nano development boards) are used, and trending development environments, languages (Python, C++), and libraries (Numpy, Pandas, Matplotlib, SciKits) are implied. The inclusion of the presentation of these practical applications aims to increase the students' confidence that the knowledge received increases not only their level of theoretical knowledge, but also their chances of being hired by local companies that carry out activities in the field.

The paper aims to study the applicability and limitations of the solution resulting from a design process for an intelligent system supporting people with special needs who are not physically able to control a wheelchair using classical systems. The intelligent system uses information from smart sensors and offers a control system that replaces the use of a joystick. The necessary movements of the chair in the environment can be determined by an intelligent vision system analyzing the direction of the patient's gaze and point of view, as well as the actions of the head. In this approach, an important task is to detect the destination target in the 3D workspace. This solution has been evaluated, outdoor and indoor, under different lighting conditions. In order to design the intelligent wheelchair, and because sometimes people with special needs also have specific problems with their optical system (e.g., strabismus, Nystagmus) the system was tested on different subjects, some of them wearing eyeglasses. During the design process of the intelligent system, all the tests involving human subjects were performed in accordance with specific rules of medical security and ethics. In this sense, the process was supervised by a company specialized in health activities that involve people with special needs. The main results and findings are as follows: validation of the proposed solution for all indoor lightning conditions; methodology to create personal profiles, used to improve the HMI efficiency and to adapt it to each subject needs; a primary evaluation and validation for the use of personal profiles in real life, indoor conditions. The conclusion is that the proposed solution can be used for persons who are not physically able to control a wheelchair using classical systems, having with minor vision deficiencies or major vision impairment affecting one of the eyes.

In this paper the authors theoretically evaluated the film thickness in a miniature angular contact ball bearing in grease and oil lubricated conditions and correlated with the variation of the electrical resistance experimentally determined by using tribometer CETR UMT-2. For grease the authors determined the film thickness considering base oil viscosity. The rotational speed of the inner race varied between 1 and 500 rpm and the axial load applied on a 7000C angular contact ball bearing was 8.1 N. The IVR and EHD lubrication regimes were analytically evaluated depending on the rotational speed and lubricants for both inner and outer race. Based of the dependence between film thickness and electrical resistance obtained for oil was evaluated the real film thickness obtained by using the grease. In case of oil lubricated condition, the experimental results showed an increase of electrical resistance with the rotational speed caused by increasing of the film thickness according to the theoretical model. For grease lubricated condition, ”a V-shaped pattern” was obtained. High values of electrical resistance at very low rotational speeds were observed. By increasing the rotational speed the electrical resistance for grease decreases until a limit speed and increase continuum over this limit. This behaviour was also reported in literature by ball-disc interferometry that showed at very low speeds that the thickener is dominant and over a speed limit the base oil is dominant in generating the film thickness.

Do accountants clearly understand the benefits and challenges of using AI? Do they perceive AI as a threat? The adoption of AI in the accounting field has increased significantly in the last few years. Since the techniques continue to evolve, more companies will integrate these solutions to facilitate the accounting processes. Therefore, the accountants’ skills should be adapted to efficiently use these solutions and continue to provide valuable support. This study explores the perception of accounting practitioners regarding the most important benefits and challenges of using AI-based technologies and analyses whether AI is being perceived as a threat that might impact employability. The data were collected during June–August 2021 using a questionnaire addressed to accounting practitioners from Romania. The exploratory research was conducted by statistically analysing the data collected. The results highlight that the practitioners have a clear understanding regarding the main benefits and challenges associated with the use of AI-based solutions in accounting processes, and AI is not perceived as a threat to employability; however, practitioners acknowledge that skills transformation is required and are willing to undergo the changes. By providing a glimpse of the main drivers that encourage accounting practitioners to embrace AI, employers, professional bodies and academia can address the main concerns and continue to support the practitioners in adapting their skills.