Automatic Adaptation Research Articles

Single‐Shot Auto‐Exposure and High Dynamic Range Imaging Based on Asynchronous Operation of Pixel Array Circuitry

ABSTRACTThis paper introduces an image formation technique that realizes automatic adaptation to the illumination conditions during image capture. This adaptation takes place asynchronously in a single shot, requiring neither external control nor iterations on the exposure setting. These characteristics are especially suitable for applications such as visual robot navigation, where illumination may abruptly change and immediate decisions must be made. The equations that model the proposed technique show exposure‐time independence. Moreover, each pixel adapts itself according to the ratio between local and global illuminations. The dependence on this ratio—which is a relative rather than absolute magnitude—along with the aforementioned independence on exposure time means that, theoretically, any illumination scenario can be represented within an arbitrary signal range. We also address the CMOS implementation of this technique. An extensive set of electrical simulations confirms its auto‐exposure and high‐dynamic‐range encoding capabilities. In particular, we simulated the capture of a 118‐dB scene on a ‐px array over five orders of magnitude of ambient illumination. Corner and mismatch simulations reveal strong robustness of the proposed circuitry against fabrication non‐idealities. Finally, although we aim to applications in which large image resolution is not a critical specification, strategies to minimize the impact of the required focal‐plane circuit elements on the pixel pitch are examined.

Automatic preload adaptation for rack-and-pinion drives to maximize performance and energy efficiency

AbstractElectrically preloaded rack-and-pinion drives are typically utilized in machine tools to precisely move heavy loads over long travel distances. In the state of the art, the preload torque is commonly set to a constant value during commissioning. This poses a conflict of objectives between maximizing performance and minimizing energy consumption. To resolve this, this publication presents a novel approach to automatically adapt the preload torque during operation. For this purpose, the established preload control is extended by a simple control law that adapts the preload torque to the current operating state within the permissible limits. This way, higher preload torques are only applied, if the resulting higher system stiffness is beneficial for the drive performance. Otherwise, the preload torque is reduced to save energy. The automatic preload adaptation is experimentally validated on a test system with industry standard components. This involves both system-theoretical analyses and practical test scenarios.

A stable method for task priority adaptation in quadratic programming via reinforcement learning

In emerging manufacturing facilities, robots must enhance their flexibility. They are expected to perform complex jobs, showing different behaviors on the need, all within unstructured environments, and without requiring reprogramming or setup adjustments. To address this challenge, we introduce the A3CQP, a non-strict hierarchical Quadratic Programming (QP) controller. It seamlessly combines both motion and interaction functionalities, with priorities dynamically and autonomously adapted through a Reinforcement Learning-based adaptation module. This module utilizes the Asynchronous Advantage Actor–Critic algorithm (A3C) to ensure rapid convergence and stable training within continuous action and observation spaces. The experimental validation, involving a collaborative peg-in-hole assembly and the polishing of a wooden plate, demonstrates the effectiveness of the proposed solution in terms of its automatic adaptability, responsiveness, flexibility, and safety.

Illustrating the benefits of efficient creation and adaption of behavior models in intelligent Digital Twins over the machine life cycle

The concept of the Digital Twin, which in the context of this paper is the virtual representation of a production system or its components, can be used as a "digital playground" to master the increasing complexity of these assets. One of the central subcomponents of the Digital Twin are behavior models that can enable benefits over the entire machine life cycle. However, the creation, adaption and use of behavior models throughout the machine life cycle is very time-consuming, which is why approaches to improve the cost-benefit ratio are needed. Furthermore, there is a lack of specific use cases that illustrate the application and added benefit of behavior models over the machine life cycle, which is why the universal application of behavior models in industry is still lacking compared to research. This paper first presents the fundamentals, challenges and related work on Digital Twins and behavior models in the context of the machine life cycle. Then, concepts for low-effort creation and automatic adaption of Digital Twins are presented, with a focus on behavior models. Finally, the aforementioned gap between research and industry is addressed by demonstrating various realized use cases over the machine life cycle, in which the advantages as well as the application of behavior models in the different life cycle phases are shown.

Multi agent collaborative search algorithm with adaptive weights

AbstractThis paper presents a new version of Multi Agent Collaborative Search (MACS) with Adaptive Weights (named MACS‐AW). MACS is a multi‐agent memetic scheme for multi‐objective optimization originally developed to mix local and population‐based search. MACS was proven to perform well on a number of test cases but had three limitations: (i) the amount of computational resources allocated to each agent was not proportional to the difficulty of the sub‐problem the agent had to solve; (ii) the population‐based search (called social actions in the following) was using only one differential evolution (DE) operator with fixed parameters; (iii) the descent directions were not adapted during convergence, leading to a loss of diversity. In this paper, we propose an improved version of MACS, that implements: (i) a new utility function to better manage computational resources; (ii) new social actions with multiple adaptive DE operators; (iii) an automatic adaptation of the descent directions with an innovative trigger to initiate adaptation. First, MACS‐AW is compared against some state‐of‐art algorithms and its predecessor MACS2.1 on some standard benchmarks. Then, MACS‐AW is applied to the solution of two real‐life optimization problems and compared against MACS2.1. It will be shown that MACS‐AW produces competitive results on most test cases analysed in this paper. On the standard benchmark test set, MACS‐AW outperforms all other algorithms in 11 out of 30 cases and comes second in other 8 cases. On the two real engineering test set, MACS‐AW and its predecessor obtain same results.

Automatic Learning Rate Adaption for Memristive Deep Learning Systems.

As a possible device to further enhance the performance of the hybrid complementary metal oxide semiconductor (CMOS) technology in the hardware, the memristor has attracted widespread attention in implementing efficient and compact deep learning (DL) systems. In this study, an automatic learning rate tuning method for memristive DL systems is presented. Memristive devices are utilized to adjust the adaptive learning rate in deep neural networks (DNNs). The speed of the learning rate adaptation process is fast at first and then becomes slow, which consist of the memristance or conductance adjustment process of the memristors. As a result, no manual tuning of learning rates is required in the adaptive back propagation (BP) algorithm. While cycle-to-cycle and device-to-device variations could be a significant issue in memristive DL systems, the proposed method appears robust to noisy gradients, various architectures, and different datasets. Moreover, fuzzy control methods for adaptive learning are presented for pattern recognition, such that the over-fitting issue can be well addressed. To our best knowledge, this is the first memristive DL system using an adaptive learning rate for image recognition. Another highlight of the presented memristive adaptive DL system is that quantized neural network architecture is utilized, and there is therefore a significant increase in the training efficiency, without the loss of testing accuracy.

Investigating intermediate shapes for multi-stage forming of cranial implants

Single Point Incremental Forming (SPIF), a numerically controlled sheet forming technique, excels in adaptability and cost reduction for small-scale production and customized parts. This study explores multi-stage SPIF, focusing on applications in the medical domain, particularly the fabrication of personalized medical implants. Addressing critical challenges related to formability and accuracy, this study emphasizes the significance of multi-stage forming and the importance of the design of intermediate shapes. Forming a cranial implant using one intermediate stage is studied, preliminary in pure Zinc as a cost-effective alternative for heated Ti. The intermediate shape is generated using automatic geometry adaptations such as translating and scaling the CAD model. Nine different experiments are conducted and compared in terms of final geometric accuracy and thickness distributions. The study shows that the proposed intermediate shapes for multi-stage forming significantly improve the geometric accuracy from 1.08 mm mean absolute deviation in traditional single stage forming to 0.41–0.59 mm. However, the used intermediate shapes have a strong but varying effect on thickness distributions throughout the part, showing the importance and sensitivity of the exact intermediate geometry. These insights provide a valuable understanding of the process, shaping the pathway for enhanced utilization of SPIF and setting the stage for future improvements.

Design and development of a feedback system for automatic treadmill speed adaptation

Treadmills with automatic speed adjustment offer a unique advantage: they can mimic outdoor conditions for specific training or testing protocols. This versatility makes them highly applicable in both sports and rehabilitation settings. This study presents a novel framework based on a feedback control loop system, conceived to control a treadmill belt speed using a non-invasive and low-cost sensor (Microsoft Kinect) to detect the users’ position and avoid object obstruction issues. The speed of the treadmill belt is regulated according to the user’s speed, by means of a proportional-integrative-derivative (PID) controller and a parabolic gain function. By tuning the gain function parameters, the user can customize the response of the treadmill. Position data collected during exercise using the Microsoft Kinect sensor was compared with that collected with a stereophotogrammetric motion capture system, showing promising results in terms of accuracy in position assessment. The comparison highlighted a 0.9 degree of correlation between the two systems during the running and cross-country indoor skiing tests performed. In addition, upon considering the relationship between the differences and averages of the two measures, no systematic bias was identified. The system proved to be functional for running and cross-country skiing, and it can therefore re-create similar typical characteristics of outdoor environments (e.g., speed and slope) thanks to the non-invasive user’s position detection and the customizable gain function.

ОСОБЛИВОСТІ ВИКОРИСТАННЯ СУЧАСНИХ ІНФОРМАЦІЙНОКОМУНІКАЦІЙНИХ ТЕХНОЛОГІЙ У ВИЩІЙ МАТЕМАТИЧНІЙ ОСВІТІ

The article discusses the relevance of using information and communication technologies (ICT) in higher mathematical education within the context of the contemporary educational environment characterized by rapid technological changes. It is noted that effective utilization of ICT, such as electronic journals, multimedia technologies, testing systems, and computer mathematics systems, contributes to the improvement of the teaching and learning process. Emphasis is placed on the necessity of reviewing and modifying teaching methodologies for mathematical disciplines to integrate innovative solutions. Furthermore, the use of information and communication technologies in higher mathematical education is deemed highly relevant due to the rapid development of the modern technological environment and the needs of the contemporary job market. The implementation of electronic journals, multimedia technologies, testing systems, and computer mathematics not only enhances the efficiency of teaching mathematical disciplines but also helps create a solid foundation for the future professional success of students. Electronic journals serve not only as convenient tools for administrative processes but also as platforms for active interaction among students, teachers, and administration, fostering the development of academic community and stimulating discussions and collaboration. Multimedia technologies, such as video lessons, interactive exercises, and simulations, provide students with opportunities to better grasp material and understand complex concepts, opening up new possibilities for personalized learning and adaptation to various learning styles. Testing systems, particularly those based on interactive platforms, enable instructors to effectively monitor students' knowledge and track their progress, as well as provide opportunities for personalized learning through automatic adaptation of task complexity Computer mathematics systems, including software packages for numerical and symbolic computations, become an integral part of modern mathematical education, empowering students to efficiently solve complex mathematical problems and explore new concepts, thus stimulating their creative potential and innovative approach to problem-solving. Together with educational platforms and distance learning systems, these tools contribute to the modernization and optimization of the educational process, leading to the improvement of the quality of higher mathematical education and the preparation of qualified professionals for the contemporary world.

Automatic Real-Time Adaptation of TrainingSession Difficulty Using Rules andReinforcement Learning in the AI-VT ITS

Automatic Real-Time Adaptation of TrainingSession Difficulty Using Rules andReinforcement Learning in the AI-VT ITS

Engineering Water‐Stiffening Polymers via PEG‐Sidechain‐Mediated Microphase Separation

AbstractStiffness‐variable polymers have found their enormous potential in practical applications that require automatic adaptation and modulation. However, prevailing materials are often constrained by specific operational triggers, limited mechanical responsiveness, and inadequate universality. In this study, drawing inspiration from the mechanism of non‐solvent induced phase separation (NIPS), a novel series of polymers exhibiting reversible water‐triggered stiffening is designed and synthesized by covalently incorporating poly(ethylene glycol) (PEG) as a non‐solvent mediator onto the hydrophobic poly(meth)acrylates backbones. Owing to the hydrophilicity difference between PEG sidechains and backbones, these polymers display swift and substantial alterations in stiffness, with large‐scale and regulable changes (from several‐fold to exceeding 200‐fold) upon water penetration, resulting from water‐induced microphase separation. Meanwhile, the polymers also possess intrinsic dual‐responsive shape‐memory properties. Due to the excellent commercial availability, versatility, and designability of these polymers, the work provides novel perspectives for the advancement of water‐triggered stiffening materials and opens avenues for their prospective applications in various domains.

Piano harmony automatic adaptation system based on deep reinforcement learning

Piano harmony is a combination of multiple notes, which plays a vital role in enriching the expression of melody. However, research on the piano harmony automatic allocation system is insufficient, and the practical application of reinforcement learning technology is scarce. Based on this, this paper determines the key of the piano harmony automatic allocation system through the in-depth analysis of piano harmony, then arranges the piano harmony with the help of reinforcement learning rather than manual methods and designs a complete piano harmony allocation system it provides powerful help for piano music creation. Sufficient experiments show that the piano harmony automatic orchestration system based on reinforcement learning proposed in this paper has high accuracy, and the orchestration effect has a similarity of 93% with the creation of musicians, which the majority of piano scholars have recognized.

Driving toward Connectivity: Vehicular Visible Light Communications Receiver with Adaptive Field of View for Enhanced Noise Resilience and Mobility.

Wireless communication represents the basis for the next generation of vehicle safety systems, whereas visible light communication (VLC) is one of the most suitable technologies for this purpose. In this context, this work introduces a novel VLC receiver architecture that integrates a field-of-view (FoV) adaptation mechanism in accordance with the optical noise generated by the sun. In order to demonstrate the benefits of this concept, a VLC prototype was experimentally tested in an infrastructure-to-vehicle (I2V) VLC configuration, which uses an LED traffic light as the transmitter. At the receiver side, an automatic FoV adaptation mechanism was designed based on a mechanical iris placed in front of a photodetector. Adjustments were made based on the values recorded by a multi-angle light sensor, built with an array of IR photodiodes covering an elevation from 0° to 30° and an azimuth from -30° to 30°. Depending on the incidence of solar light, the mechanical iris can adjust the FoV from ±1° to ±22°, taking into account both the light irradiance and the sun's position relative to the VLC receiver. For experimental testing, two identical VLC receivers were used: one with an automatic FoV adjustment, and the other with a ±22° fixed FoV. The test results performed at a distance of 50 m, in the presence of solar irradiance reaching up to 67,000 µW/cm2, showed that the receiver with a fixed FoV saturated and lost the communication link most of the time, whereas the receiver with an adjustable FoV maintained an active link throughout the entire period, with a bit error rate (BER) of less than 10-7.

Complexity biomechanics: a case study of dragonfly wing design from constituting composite material to higher structural levels.

Presenting a novel framework for sustainable and regenerative design and development is a fundamental future need. Here we argue that a new framework, referred to as complexity biomechanics, which can be used for holistic analysis and understanding of natural mechanical systems, is key to fulfilling this need. We also present a roadmap for the design and development of intelligent and complex engineering materials, mechanisms, structures, systems, and processes capable of automatic adaptation and self-organization in response to ever-changing environments. We apply complexity biomechanics to elucidate how the different structural components of a complex biological system as dragonfly wings, from ultrastructure of the cuticle, the constituting bio-composite material of the wing, to higher structural levels, collaboratively contribute to the functionality of the entire wing system. This framework not only proposes a paradigm shift in understanding and drawing inspiration from natural systems but also holds potential applications in various domains, including materials science and engineering, biomechanics, biomimetics, bionics, and engineering biology.

Concept-drifts adaptation for machine learning EEG epilepsy seizure prediction.

Seizure prediction remains a challenge, with approximately 30% of patients unresponsive to conventional treatments. Addressing this issue is crucial for improving patients' quality of life, as timely intervention can mitigate the impact of seizures. In this research field, it is critical to identify the preictal interval, the transition from regular brain activity to a seizure. While previous studies have explored various Electroencephalogram (EEG) based methodologies for prediction, few have been clinically applicable. Recent studies have underlined the dynamic nature of EEG data, characterised by data changes with time, known as concept drifts, highlighting the need for automated methods to detect and adapt to these changes. In this study, we investigate the effectiveness of automatic concept drift adaptation methods in seizure prediction. Three patient-specific seizure prediction approaches with a 10-minute prediction horizon are compared: a seizure prediction algorithm incorporating a window adjustment method by optimising performance with Support Vector Machines (Backwards-Landmark Window), a seizure prediction algorithm incorporating a data-batch (seizures) selection method using a logistic regression (Seizure-batch Regression), and a seizure prediction algorithm with a dynamic integration of classifiers (Dynamic Weighted Ensemble). These methods incorporate a retraining process after each seizure and use a combination of univariate linear features and SVM classifiers. The Firing Power was used as a post-processing technique to generate alarms before seizures. These methodologies were compared with a control approach based on the typical machine learning pipeline, considering a group of 37 patients with Temporal Lobe Epilepsy from the EPILEPSIAE database. The best-performing approach (Backwards-Landmark Window) achieved results of 0.75 ± 0.33 for sensitivity and 1.03 ± 1.00 for false positive rate per hour. This new strategy performed above chance for 89% of patients with the surrogate predictor, whereas the control approach only validated 46%.

Development of Fuzzy Logic Controller in Automatic Vehicle Navigation using IoT

The integration of fuzzy logic controllers in automatic vehicle navigation systems represents a significant advancement in intelligent transportation systems, especially when paired with Internet of Things (IoT) functionalities and optimized through genetic algorithms. This innovative fusion harnesses the precision of fuzzy logic, the connectivity of IoT, and the optimization capabilities of genetic algorithms to transform automatic vehicle navigation. Fuzzy logic controllers excel in managing uncertainty and imprecision, providing decision-making capabilities akin to human reasoning. By simultaneously assessing multiple inputs and determining actions based on degrees of truth, fuzzy logic enables safe and efficient navigation in dynamic driving environments with fluctuating variables like obstacle proximity and traffic flow. IoT integration enhances navigation systems by enabling real-time data collection and sharing among vehicles and infrastructure, fostering adaptive route planning and improving the overall navigation experience. Genetic algorithms further optimize system performance by iteratively adjusting fuzzy logic controller parameters, ensuring efficient decision-making tailored to specific performance criteria such as travel time and fuel consumption. This collaborative integration of fuzzy logic controllers, IoT, and genetic algorithms offers a holistic solution to the challenges of automatic vehicle navigation, enhancing safety, efficiency, and adaptability in complex driving scenarios. Beyond enhancing individual vehicle performance, this approach contributes to overall transportation system efficiency and safety by mitigating traffic congestion, reducing emissions, and minimizing accidents. Consequently, these integrated systems address crucial societal challenges and pave the way for widespread adoption of autonomous vehicles in the future.

A Self-Adaptive Automatic Incident Detection System for Road Surveillance Based on Deep Learning.

We present an automatic road incident detector characterised by a low computational complexity for easy implementation in affordable devices, automatic adaptability to changes in scenery and road conditions, and automatic detection of the most common incidents (vehicles with abnormal speed, pedestrians or objects falling on the road, vehicles stopped on the shoulder, and detection of kamikaze vehicles). To achieve these goals, different tasks have been addressed: lane segmentation, identification of traffic directions, and elimination of unnecessary objects in the foreground. The proposed system has been tested on a collection of videos recorded in real scenarios with real traffic, including areas with different lighting. Self-adaptability (plug and play) to different scenarios has been tested using videos with significant scene changes. The achieved system can process a minimum of 80 video frames within the camera's field of view, covering a distance of 400 m, all within a span of 12 s. This capability ensures that vehicles travelling at speeds of 120 km/h are seamlessly detected with more than enough margin. Additionally, our analysis has revealed a substantial improvement in incident detection with respect to previous approaches. Specifically, an increase in accuracy of 2-5% in automatic mode and 2-7% in semi-automatic mode. The proposed classifier module only needs 2.3 MBytes of GPU to carry out the inference, thus allowing implementation in low-cost devices.

Post-process updating of model parameters to approximate thermal errors of machine tools operating in different configurations

Machine tool (MT) thermal errors induced by external and internal heat sources (e.g. changing environment, friction and power losses in MT components) are an important element in machined workpiece inaccuracies. In the past few decades, indirect software compensation techniques have been used to address thermal errors on account of their economic and ecological aspects. Many slightly different approaches are described in the literature. If thermal error models are properly identified based on experiments with sufficiently varying input parameters, they usually work within similar calibration and verification conditions. As the sensory equipment of machines increases, models can be adapted to regions with higher thermo-mechanical system nonlinearity and inhomogeneity through the introduction of deformation feedback from direct measurements into model structures. However, adaptive functionalities require discrete interruptions of the MT's work cycle, which threaten the integrity of the machined surface and complicate the model structure and thus also implementation and industrial deployment possibilities. In this research, a novel approach of automatic post-process adaptation suitable for repetitive manufacturing, which takes advantage of transfer functions (TF), is proposed for thermal error reduction. The method respects basic heat transfer mechanisms in the MT structure by combining linear parametric models, requires a minimum of additional gauges, and feedback from direct measurements is obtained only during technological pauses in the production process. Post-process adaptation is solved retrospectively by updating model parameters using the heuristic method of genetic algorithms (GA). First, the approach was tuned on a case study of a heavy-duty milling machine with a horizontal headstock for a configuration with an exchangeable spindle head and a configuration with a change in the position of the headstock in the workspace, both of which were different from the initial model's calibration range. Subsequently, the developed approach was applied to repeated production on a medium-sized milling centre. Another goal of the paper is to emphasize the need for quality input information for modelling efforts and the industrial applicability of scientific results.

Deep Reinforcement Learning for Automatic Run-Time Adaptation of UWB PHY Radio Settings

Deep Reinforcement Learning for Automatic Run-Time Adaptation of UWB PHY Radio Settings

Development analysis of intelligent robots in manufacturing industry

With the continuous development of the manufacturing industry, the application of intelligent robots is becoming more and more extensive. Many emerging technologies can be applied to intelligent robots. Typical intelligent robots still have room for further improvement in terms of automatic control and adaptation to the surrounding environment. And independent learning of production tasks and realization of human-computer interaction is the future development direction of industrial robots. In view of these deficiencies, the development of intelligent robots is particularly important. This paper introduces the application fields, key technologies and needs of intelligent robots, and explains the important position of intelligent robots in the future manufacturing industry. The functions of intelligent robot control, perception and human-computer interaction based on artificial intelligence technology are studied. Possible future challenges, limitations and problems will be presented at the end of this paper.