This article introduces a novel approach that combines a multimodel technique with model-free adaptive control (MFAC) to address the limitations of the full-form dynamic linearization (FFDL) method. The existing FFDL data model fails to incorporate system dynamics and disregards valuable prior knowledge. This study proposes a more robust, time-varying data model achieved by integrating a network of local autoregressive with exogenous inputs (ARX) models into the FFDL structure. This allows the new model to incorporate partial, prior knowledge of the nonlinear system’s dynamics across various operating points. By leveraging the structural similarity between FFDL and ARX models, the proposed multi-ARX (M-ARX) model provides a more global and physically grounded understanding of the system, replacing the nonphysical approximation of the traditional FFDL method. The proposed controller’s practicality is demonstrated on a nonlinear multiple-input multiple-output (MIMO) three-tank system (3TS). Its performance is compared with the traditional MFAC-FFDL approach across diverse measurement noise conditions to more accurately reflect real-world experimental scenarios, employing multiple control performance criteria. The results show that the novel M-ARX-MFAC consistently tracks desired references by incorporating local system information through weighted ARX coefficients. Its data model parameters demonstrate a more robust evolution when subjected to measurement noise. When operating with identical control parameters, M-ARX-MFAC significantly outperforms MFAC-FFDL in terms of both input intensity and smoothness across various measurement noise scenarios.

This paper proposes the integration of Bumpless Transfer Control (BTC) strategy into an Event-based Cloud Control System (EBCCS). The proposed EBCCS is a cloud-based Nonlinear Model Predictive Control (NMPC) framework equipped with a time-delay compensator, which utilizes the most recent input sequence during communication and computation delays until a new input sequence is received from the cloud controller. To minimize abrupt jumps (bumps) in the control input during the transition from the previous input sequence to the newly computed one, a novel simulation-based BTC strategy is introduced. The proposed BTC consists of two components. The first component of the proposed BTC approach initiates NMPC computation using simulated future system states, leveraging the knowledge of previously computed input sequences that will be applied during the delay. The second component, applied after NMPC computation, matches the current system states with the predicted system states and their corresponding input, optimizing the starting point of the new sequence for smoother transitions. This BTC strategy suppresses bumps during transitions between control sequences without modifying the core structure of the NMPC optimization algorithm, ensuring straightforward integration into existing control frameworks. The effectiveness of the proposed EBCCS equipped with the BTC strategy is demonstrated through simulations on a crane pendulum system and an inverted pendulum system, highlighting its effectiveness in improving control performance under time delays. The results indicate that the BTC strategy can significantly enhance stability and performance in event-based cloud control settings, paving the way for more robust implementation in real-world, delay-prone cloud control systems.

This contribution addresses vessel trajectory prediction (VTP), focusing on the evaluation of different deep learning-based approaches. The objective is to assess model performance in diverse traffic complexities and compare the reliability of the approaches. While previous VTP models overlook the specific traffic situation complexity and lack reliability assessments, this research uses a probability of detection analysis to quantify model reliability in varying traffic scenarios, thus going beyond common error distribution analyses. All models are evaluated on test samples categorized according to their traffic situation during the prediction horizon, with performance metrics and reliability estimates obtained for each category. The results of this comprehensive evaluation provide a deeper understanding of the strengths and weaknesses of the different prediction approaches, along with their reliability in terms of the prediction horizon lengths for which safe forecasts can be guaranteed. These findings can inform the development of more reliable vessel trajectory prediction approaches, enhancing safety and efficiency in future inland waterways navigation.

Using data sources beyond the Automatic Identification System (AIS) to represent the context in which a vessel is navigating and consequently improve situation awareness is still rare in machine learning approaches to vessel trajectory prediction (VTP). In inland shipping, where vessel movement is constrained within fairways, supplementary navigational context information is indispensable. In this contribution targeting inland VTP, Gaussian Mixture Models are applied, on a fused dataset of AIS and discharge measurements, to generate multi-modal distribution curves, capturing typical lateral vessel positioning in the fairway and dislocation speeds along the waterway. By subsequently sampling the probability density curves of the GMMs, feature vectors are derived which are used, together with spatio-temporal vessel features and fairway geometries, as input to a VTP transformer model. The incorporation of these distribution features of both the current and forthcoming navigation context improves prediction accuracy. The superiority of the model over a previously proposed navigation context-sensitive transformer model for inland VTP is shown. The novelty lies in the provision of preprocessed, statistics-based features representing the conditioned spatial context, rather than relying on the model to extract relevant features for the VTP task from contextual data. Oversimplification of the complexity of inland navigation patterns by assuming a single typical route or selecting specific clusters prior to model application is avoided by giving the model access to the entire distribution information. The methodology’s generalizability is demonstrated through the usage of data of three distinct river sections. It can be integrated into an interaction-aware prediction framework, where insights into the positioning of the actual vessel behavior in the overall distribution at the current location and discharge can further enhance trajectory prediction accuracy.

Chemistry 4.0 is the new era of chemical process industry, where digitalization, modularization, sustainability, and circular economy play key roles. A growing interest in the use of process data with the aim of gaining a better understanding of the production process and optimizing products can be observed. In chemical industry, data-driven models are used when the production process is too complex to be described by chemical laws. For generating data-driven models of industrial processes, many manual and time-consuming steps have to be carried out. This leads to delay in information acquisition and product optimization. Another key element to realize a modular automation framework is standardization, e.g., the Module Type Package (MTP). In this paper, a modular automation framework to optimize coating formulations with data-driven modeling in industrial environment is presented. Therefore, the presented framework automates the acquisition of data via modular and standardized interfaces in a process control system (PCS), data preparation for data-driven modeling, model training, and deployment of the model to propose new coating formulations. The framework shown in this paper is applied to optimize the scratch resistance of a coating using Gaussian processes (GPs) as data-driven modeling method. For the novel automation of machine learning driven design of experiments (DoE), the Bayesian optimization is used.

Highly automated or autonomous systems utilize machine learning-based approaches to perceive the environment and make decisions based on the obtained information. These approaches are highly depended on the model used, training data, and environmental conditions. While high performance can be expected in certain and suitable situations, unknown or uncertain situations might lead to undetected underperformance. In case of object detection, the quantification of reliability of a particular prediction is usually given by a detection score predicted by the trained model. While a higher score indicates higher confidence, it does not reflect the actual uncertainty of the prediction considering situational variations. In this paper, the predicted detection score is converted into a true-positive rate to decide whether to accept or reject a prediction. Furthermore, the detection rate is used as a function over distance to determine the dynamic performance expectations using situational knowledge and reliability requirements.

Autonomous transportation systems such as road vehicles or vessels require the consideration of the static and dynamic environment to dislocate without collision. Anticipating the behavior of an agent in a given situation is required to adequately react to it in time. Developing deep learning-based models has become the dominant approach to motion prediction recently. The social environment is often considered through a CNN-LSTM-based sub-module processing a social tensor that includes information of the past trajectory of surrounding agents. For the proposed transformer-based trajectory prediction model, an alternative, computationally more efficient social tensor definition and processing is suggested. It considers the interdependencies between target and surrounding agents at each time step directly instead of relying on information of last hidden LSTM states of individually processed agents. A transformer-based sub-module, the Social Tensor Transformer, is integrated into the overall prediction model. It is responsible for enriching the target agent's dislocation features with social interaction information obtained from the social tensor. For the awareness of spatial limitations, dislocation features are defined in relation to the navigable area. This replaces additional, computationally expensive map processing sub-modules. An ablation study shows, that for longer prediction horizons, the deviation of the predicted trajectory from the ground truth is lower compared to a spatially and socially agnostic model. Even if the performance gain from a spatial-only to a spatial and social context-sensitive model is small in terms of common error measures, by visualizing the results it can be shown that the proposed model in fact is able to predict reactions to surrounding agents and explicitely allows an interpretable behavior.

Electric vehicles (EVs) are promising alternatives to carbonized propulsion-based vehicles. They are capable of reducing environmental degradation without compromising driving performance. Power management strategies (PMS) are particularly essential for electrified vehicles to ensure optimal power split between on-board energy storage sources and to meet operational requirements of each source. However, optimization concept in PMS, have been constantly addressed in literature to achieve optimal power handling decisions in real-time, particularly under unknown driving conditions. In this contribution, an intelligent rule-based PMS with embedded offline-optimized control parameters and online driving state recognition is proposed to achieve optimal power handling decisions for EVs situatively and adaptively. A set of characteristic variables defining driving states have been extracted from representative segments of several driving cycles, to which optimized control strategies are tuned offline. Three different driving cycles representing urban, highway, and mixed trip conditions have been implemented for comparative investigation of achieved results. The analysis of results reveals the potential of proposed PMS to reduce the energy consumption by 13.6 – 30.9 %.

Advanced Driver Assistance Systems (ADAS) are systems developed to assist the human driver and therefore to make driving safer and better. Understanding and predicting human driving behavior play an important role in the development of assistance systems. In this contribution, the development of a driver assistance system is based on the prediction of driving behaviors. The driving patterns of three different behaviors are modeled including left/right lane change and lane keeping. A driving simulator is used to simulate a highway scene. The implementation of a prediction system based on different machine learning approaches such as Hidden Markov Model (HMM), Support Vector Machine (SVM), Convolutional neural networks (CNNs), and Random Forest (RF) is accomplished. In addition, eye-tracking information is integrated. The task is to predict behaviors based on the measurement. As test, a 10-fold cross-validation is used based on data sets from driving simulator and applied to compare the performance of different algorithms. In combination with related results in terms of accuracy (ACC), detection rate (DR), and false alarm rate (FAR), the performance and effectiveness of the developed prediction systems are evaluated. The results show that the performance of RF algorithm is the best of all four algorithms compared. Combining environmental and eye-tracking data the RF algorithm achieved the best results. All ACC values are larger than 99 %. Afterwards, two RF-based prediction models with and without eye-tracking data are developed for online test. Finally, some application samples are suggested for driver assistance. The results calculated by the proposed model are shown on a user interface to help the drivers to see when it is suitable to turn left, to turn right, or to keep the direction.

The situation awareness of drivers during takeover from autonomous to manual mode is important for avoidance of accidents. Previous studies have revealed that takeover time and general performance vary strongly in different situations. The studies also revealed that the variation is due to surrounding traffic conditions, complexity of the driving scenario, secondary tasks, speed of ego vehicle, and takeover request experience. The aim of this study is to further explore the scope and dependencies of the aforementioned variables to better equip driver assistance and supervision systems with the necessary framework to suitably assist drivers during takeovers. In other words, the intention is to define a formal set of rules to enable the automated driving system determine a suitable takeover request time for different scenarios. First, this contribution discusses the design of takeover variables such that the effects of the variables are systematically varied to generate different driving situations. Afterwards, experimental results under different variable combinations are discussed. The results include a comparison of objective measures and subjective measures. An initial set of rules are established to model the interaction to improve intelligent driver assistance systems.

Partially automated driving releases the driver from controlling the vehicle, but the driver still needs to monitor the driving situation steadily and be able to take over the control in critical situations when needed. In such a situation, problems arise as to how and when should the driver be informed and how to make sure the driver understands the general situation and specifically also the intention of the system correctly. Driver-vehicle interfaces (DVIs) present the necessary information to the driver. Many studies focusing on the time for Takeover-request (TOR) and the takeover time of the driver are performed leading to various results. The question: “Can a general conclusion about upper and lower limits of these two variables be drawn, which is applicable for all drivers in all situations?” should be answered. In the situation in which the driver should take over the control of the vehicle, multiple driving modes are involved. Mode error or confusion may happen when the vehicle has multiple modes. This leads to another question to be discussed in this contribution: Is it possible to use one DVI to show the status of several autonomous driving levels and corresponding functionalities to avoid mode error or confusion? To answer these questions, two DVIs are designed considering multiple autonomous driving levels and their functionalities to increase the situation/mode awareness of the driver. A simulator study with 38 participants was conducted to study the takeover time and behavior of drivers under different critical situations as well as to compare the usability. Results show that the takeover time depends upon the sequence of experiencing the TOR and the complexity of the takeover situation. Furthermore, two proposed interfaces could contribute to increasing the situation awareness of the driver. The limits and trends of developing the TOR are concluded based on the literature review and the experiments performed. It is concluded that using an average TOR time for every individual driver in all critical situations is not appropriate. Discussion with respect to TOR to ensure a safe takeover regarding non-driving-related tasks and takeover experiences is derived from the performed experiments.

Research and development of human driving behaviors play an important role in the development of assistance systems. In this contribution, a driving behaviors prediction model is based on a newly developed approach combining different Hidden Markov Models (HMM) cooperatively combined by Fuzzy Logic (FL). Due to variations of individual human drivers decision behavior the task to classify related behaviors based on individually trained models is difficult. The FL approach will be used for additional distinction of driving scenes into very safe, safe, and dangerous driving scenarios. For each scenario corresponding HMMs will be trained. Three different driving behaviors including left/right lane change and lane keeping are modelled as hidden states for the HMM. Based on observations, the algorithm calculates the most possible driving behaviors through the observation sequences. Furthermore, the observed sequences are also used for training of HMM during modeling process. To improve the prediction performance of the model, a prefilter is proposed to quantize the collected signals into observed sequences with specific features. To optimize the model performance NSGA-II was used to define the optimal thresholds of FL and the optimal prefilters of HMMs. Using experimental data from real human driving behaviors (taken from driving simulator) it can be concluded that selecting optimal thresholds will increase the performance of driving behaviors prediction. The effectiveness of the suggested fuzzy-based HMM has been successfully proved based on experiments.

Research and development of predicting driving behaviors play an important role in the development of Advanced Driver Assistance Systems (ADAS) for assisting drivers. In this contribution, an approach is developed based on Hidden Markov Model (HMM) for predicting human driving behaviors. Three different driving maneuvers including left/right lane change and lane keeping are modeled as hidden states for the HMM. Based on observations (training), the HMM approach is able to calculate the most possible driving behaviors using observed sequences. Furthermore, the observed sequences are also used for training of HMM in the modeling process. To improve the prediction performance of the model, a prefilter is proposed to quantize the collected signals into observed sequences with specific features. In this contribution the definition of a suitable prefilter will be discussed and finally optimized. The approach focuses on the definition of optimal prefilters. Here optimality is defined as the optimal segments describing a quantized prefilter mapping the vehicle's environment to quantized states. In combination with related HMM-based results in terms of accuracy, detection, and false alarm rates an optimal parameter set of the prefilter can be determined. Using experimental data from real human driving behaviors (taken from driving simulator) it can be concluded that the optimal definition of the prefilter can increase the detection rate and accuracy, and in the meanwhile decrease the false alarm rate. The effectiveness of driving behaviors prediction has been successfully proved by comparison with other methods in this contribution.

Autonomous and semi-autonomous aerial systems (AES) are vehicles which have to perform tasks in complex and dynamic environment, for example in logistics and transportation applications. In this context vehicle's behavior has to be verifiable safe. Here safe behavior denotes interaction with the environment like for example spacial movements with freedom of unacceptable risk. Traditionally, behavioral safety aspects are combined with mission-related tasks. The consequence may be unmanageable vehicle's complexity as well as unpredictable effects during the interaction with the environment leading to inapplicability of traditional safety assurance methods. Furthermore, real-time vehicle's hardware reliability assessment and management are often not considered. This paper introduces a novel simplified real-time environmental situation risk assessment approach for determination of required situational vehicle's reliability. Furthermore, the novel real-time vehicle's hardware reliability assessment and situational hardware structure and behavior control for vehicle's safe behavior assurance is presented. The introduced approach can verify vehicle's safe situational behavior by real-time environmental risk assessment and vehicle's hardware structure and emergency behavior control to minimize the situational risk. The technical proof of concept demonstrates for the first time the successful use of the introduced approach based on AES example using Virtual Robot Experimental Platform and Programmable Logic Controller.

Advanced Driving Assistance Systems could improve driving safety and comfort by supporting drivers in their driving task. To realize intelligent assistance, driver behavior prediction and recognition is an important challenge. Therefore, the aim of this study is to develop a method to predict whether a vehicle, having entered a roundabout, will choose an upcoming exit or stay within the roundabout. A field study has been conducted to collect driving behavior data for analyzing and modeling human driver behavior in interaction with roundabouts. Support vector machines proved to be a robust and efficient classification method for the roundabout leaving/ staying pattern recognition problem. From the experimental results the vehicles position can be estimated, for which the prediction becomes reliable. The steering wheel angle and angle velocity also proved to be able to provide sufficient information to predict the driver behavior at the investigated roundabouts.

Autonomous and semi-autonomous aerial systems (AES) are often needed to perform tasks in complex and dynamic environments. The safe navigation assurance as well as safety assurance of AES are open research issues. Traditional combination of safety aspects with mission related tasks and in consequence unmanageable AES system complexity as well as unpredictable effects during the spatial environment interaction makes traditional safety assurance methods inapplicable. This paper introduces Strictly Formalized Situation-Operator-Modeling (sf-SOM) technique for AES safe behavior assurance. In combination with the System Safety Surveillance and Control (SSSC) system concept a AES fall-back layer concept can be realized. In comparison to other approaches, in this concept a separation between regular behavior generating mission-tasks and safety assurance non-mission tasks is used. Furthermore, the system is separated in well-defined, safety task-specific modules and can be realized using standardized industrial programming languages and programmable safety device. Proof of concept using an industrial Programmable Logic Controller demonstrates the successful use of SSSC-based fall-back layer also for comparable applications.

Autonomous or semi-autonomous systems (AS) are needed in different application domains to simplify human tasks. Autonomous and semi-autonomous aerial systems (AES) are most complex examples due to the challenge to perform tasks in complex and dynamic environment, for example in search and rescue applications. The safe navigation assurance as well as safety assurance of AS are open research issues. Traditional combination of safety aspects with mission related tasks and in consequence unmanageable AS system complexity as well as unpredictable effects during the spatial environment interaction makes traditional safety assurance methods inapplicable. This paper presents a brief literature review concerning AES systems. A related novel concept for System Safety Surveillance and Control (SSSC) system as AS fall-back layer is introduced considering safe situational behavior, system malfunctions, and technical fall-back layer. This system is separated in well-defined, safety task-specific modules. In comparison to other approaches, in this paper safety is achieved by separation between regular behavior generation and safety assurance by emergency behavior integration and realization. Universally concept design permits the fall-back layer realization also for other applications. In this contribution a first proof of concept of SSSC-based fall-back layer realization using an experimental example is given.

This paper investigates feature selection to discard irrelevant features for dimensionality reduction and improving situation recognition process. A situation illustrating the internal structure of a system state and its related environment is based on a large set of characteristics (features). Real-time situation recognition is still a challenge because of dealing with incremental knowledge as well as imprecise, uncertain, and redundant data (features). Investigation of relevant and key situations features could effectively enhance the situation recognition performance in terms of accuracy and computational complexity. In this paper, Case-Based Reasoning (CBR) as a problem solving approach is used for situation recognition. A fuzzy SOM-based approach by integration of Situation-Operator Modeling (SOM) and Fuzzy Logic (FL) is provided for knowledge representation in CBR process. A feature selection is realized using Rough Set Theory (RST) for data mining and uncertainty management in real-time applications. Different feature selection algorithms based on RST are applied to fuzzy SOM-based CBR. An analysis of the performance of all resulting combinations is done in terms of feature reduction and situation recognition. Finally, the proposed CBR approach is realized using experiments based on driving maneuvers conducted by a professional driving simulator. This application shows the effectiveness as well as the accuracy of the introduced approach.

In driver-vehicle systems, the Human-Machine Interface (HMI) plays an important role to assist the drivers in understanding the status of the vehicle and also serves as an input device. The HMI is an essential part in different levels of autonomous driving. As long as the interaction between the driver and vehicle remains, the outputs from the assistance system should be displayed to the driver. Studying the state of the art of the displays in vehicles, it can be stated that most of the displays show the current state of the vehicle well and detailed. To drive efficiently or to make a right decision in time, the human driver has to be assisted more specifically. Increasing driving efficiency, namely decreasing fuel consumption is one of the issues in modern automotive technology. Based on the proposed concept, actions relevant to the predefined or given goals can be displayed based on Augmented Reality (AR) in a suitable manner. In such way, the driver-vehicle-loop is closed by the interface. In this contribution, three proposed interfaces sharing the same optimal strategy behind are proposed to increase the fuel efficiency. A driving simulator is used to test the proposed interfaces. It is connected to a HiL (Hardware in Loop) test rig, which simulates the powertrain of a hybrid electric vehicle (HEV). Experiments were done to validate these three interfaces. Results show that the fuel efficiency is increased by using the proposed interfaces. Furthermore, results show also that more economic driving requires more cognitive workload.

With respect to planning capabilities, a model is developed to be used for the evaluation of human operator's cognitive performance. Plans are goal-directed interaction sequences cognitively simulated before appearing in dynamic human-machine systems. In this context, human operator's actions have long-term effects as well as different consequences depending on the time of their activation. The proposed model allows the evaluation of cognitive task performance by comparing implemented decisions to the available options represented by plans. In this paper, situation operator modeling is applied to describe operator's behavior and the planning process based on the system modeled using colored Petri Nets. The Petri Net is applied for the prediction of future system states. The planning model uses a set of rules that describe the normative behavior of operators as definitions of problems and actions required to avoid these. The rules are derived systematically from the operator's objectives and from the available actions. The proposed method allows the calculation and, therefore, the prediction of future states as well as possible upcoming conflicts within the system. Corresponding rules may modify and, therefore, affect the generated plan. The method is demonstrated using a simplified air traffic control simulation. This method extends earlier methods published by the authors for the determination of available options based on Petri Nets. It can be concluded that the calculation of sequences of actions for interaction with dynamic systems with constant change over time is possible and realized for the first time.

Disaster or emergency management becomes more and more important due to an increasing number of catastrophes in combination with advanced complexity of human living, working, and society. In combination with increased options for intervention, the need for related communication and coordination support increases. System allows a suitable management in combination with skilled and experienced human operators, a flexible and situative reaction leading to a suitable guidance of reaction forces as well as of groups of human etc.

With the increasing of automotive technology, the focus of vehicle developing has been changed from former basic needs of transportation to the additionally and enhanced requirements regarding to safety and efficiency. The cause of accidents is manifold. The inappropriate driving is one of the reasons. Additional effort understanding the driver's intention in combination with extended sensor information allowing to match driver's intention and real available options will be useful with respect to warning and intervention. Based on the Situation-Operator-Modeling (SOM), the intention of the driver can be detected. Once the intended action includes dangerous element, a warning will be provided to avoid the action really to be carried out. The other reason of accident is incomplete understanding about the available choices. Stimulated by daily life, the options can also be provided to the human driver during driving. Comparing the available choices and corresponding consequences, the suitable and efficient decision can be made. This contribution introduces a novel concept about providing warnings related to danger or wrong decision based on the calculation of intended actions and options to the driver to realize a novel online driver supervision system.

Modeling of human interactions allows consideration of the different human interaction qualities to a cooperative teamwork with technical systems. The analysis of the performance of the cooperative interacting team members has become an important focus of research in recent years. The proposed approach can be used to model the behaviors for collaborations. This formalizable aspect of the teamwork uses a variety of descriptions. This contribution introduces the Situation-Operator-Modeling (SOM) approach theoretically and applies the approach with respect to the shared mutual situation awareness. Possible functionalities of SOM-based teamwork supervision and assistance are given with examples like teamwork coordination, agent learning, cognitive supervising, and automatic error detection of human errors.

Computer supported cooperative work requires a detailed understanding of the underlying working tasks and their internal relations. A specific kind of this work related knowledge can be used supporting the interactions between the interacting partners, like human operators. In addition to the working tasks, the `knowledge behind and the rules of interaction' between the cooperating units has to be formalized to be used for support. This contribution describes a new game-theoretical approach used for modeling the goal-oriented interaction and cooperation of human operators. The introduced approach is based on a suitable formalization using Situation-Operator-Modeling (SOM) as a modeling technique formalizing the interaction of systems. The approach is illustrated with examples detailing different aspects: i) semi-automated molding processes (guided human-machine-interaction), ii) air traffic management (multi-human-machine-interaction), iii) as illustrative conceptional example cooperative realization of documents (multi-human-machine-interaction). As a result it can be stated that the game-theoretic approach allows a new and specific view for the modeling of the interactions between the cooperating units. The game-theoretical view to the SOM-based structured interaction gives a suitable perspective to the strategies of possibly processible action spaces. Therefore technical understanding of human intentions and also realization of goal-oriented computer supported assistance functions can be achieved.

The estimation of state variables plays a significant role in the control of ship-mounted elastic cranes. Elastic-ship mounted crane can show very complex dynamic vibrations during their operations. A vision sensor (i.e. camera) realizing a contactless measurement sensor can be used to measure the deformations. Unfortunately effects like limited accuracy and time delay occur, which are the main inherent problems of vision sensors. The effects and related compensation approaches are studied in this work. The main goal of the work is to develop an approach to combine suitable measurement devices easy to realize with improved reliability. The task to be solved is to combine the estimations of the variable gain observer (based on vision measurements) with those of the variable gain observer (based on potentiometers). Realizing a multi-model approach the variable gain controller uses the estimated states and the roll angle to generate the required damping to control the system. Simulation results show that the variable gain observers can estimate the states and the unknown disturbance acting on the payload very well and the variable gain controller can reduce effectively the payload pendulations.

Future autonomous service robots are intended to operate in open and complex environments. This in turn implies complications ensuring safe operation. The tenor of few available investigations is the need for dynamically assessing operational risks. Furthermore, a new kind of hazards being implicated by the robot's capability to manipulate the environment occurs: hazardous environmental object interactions. One of the open questions in safety research is integrating safety knowledge into robotic systems, enabling these systems behaving safety-conscious in hazardous situations. In this paper a safety procedure is described, in which learning of safety knowledge from human demonstration is considered. Within the procedure, a task is demonstrated to the robot, which observes object-to-object relations and labels situational data as commanded by the human. Based on this data, several supervised learning techniques are evaluated used for finally extracting safety knowledge. Results indicate that Decision Trees allow interesting opportunities.

With the increasing development of complex technical systems such as power supply systems or human-machine-interaction systems, their availability, reliability and safety are always a special interest of researches. In recent years, the concept of human-centered guidance and control has been introduced into the complex technical systems to solve the growing complexity in interdependencies, uncertainties and decision making. In this contribution a human-centered approach, the Situation-Operator-Modeling (SOM), is used as an underlying approach to describe the complex interacting environmental scenes by a network of actions and states. Based on the related formal description of Human-Machine-Interaction, the representation of human behaviors, interactions, and process procedures is performed to build up a framework for the development and the implementation of optimized human-centered supervision and assistance systems. As example to the approach, a cognitive driver supervision system for lane changing maneuvers is given.

Human-centered engineering is a main concern regarding reliability issues in automation of complex processes. However, even recent automation devices are not capable to automate processes with an advanced cross linkage of process technique and manufacturing skills as occurring in traditional handcrafting processes. Analyzing handcrafting applications for automation approaches regarding human factors on the one hand and economical goals on the other hand sometimes lead to a necessity of the human worker being integrated in the process. Thus, a role mapping becomes necessary, transferring the human worker from an ideal multivariable sensory and actuatory process “element” into a process guiding and supervision role. Based on the Situation-Operator-Modeling approach, the process procedure is modeled by a set of situations. The established action space represented by a network of discrete situations is used for the implementation of a situation-based process guidance and supervision assistance system improving the situation awareness of the human worker and offering situation-based, contextual information as support for the decision making and selection process of necessary and additional available process actions. The analysis of a handcrafted molding process using No-Bake-Technique is given as an illustrative example.

Development of CSCW has been focused on cognitive cooperative agents in recent years. Modeling of human interactions allow the mapping of the human interactions qualitites using agents, supporting Cognitive Technical Systems (CTS). The proposed concept modeling can be used and integrated to the supervision and assistance of Human-Machine-Interactions by supporting the interaction using cognitive-based interfaces for communication and cooperation. The contribution introduces briefly the Situation-Operator-Modeling (SOM) approach theoretically and details on the representations with an applicable algorithm with respect to a special case of HMI: driver-vehicle interactions. The interaction between the human driver and the vehicle can be illustrated with a cyclic loop. Within the rule-based knowledge representation approach operator-specific variables are defined and identified during the interactions. Based on the experimental-based and verified experiences the concept for SOM-based CSCW is briefly developed using similar concepts of task-based HMI-modeling for group and personal assistance.

Autonomous systems are often needed to perform tasks in complex and dynamic environments. For this class of systems, traditional safety assuring methods are not satisfying due to the unknown effects of the interacting system with an open environment. Briefly speaking: What is not known during the development phase, can not be adequately considered. In order to realize a more flexible safety analysis, the internal representation of the outside world to be learned by an autonomous Cognitive Technical System, is used to identify hazardous situations. The so-called safety principles represent the hazard knowledge. These can be added to the system prior to operating time without losing the possibility of adjusting or expanding this hazard knowledge during operating time. This contribution details a new method for safety assurance and therefore proposes the introduction of so-called safety principles. Furthermore, the Cognitive Technical System provides anticipation capabilities, so that is becomes possible to expand the planning process in order to take hazard information into account. Finally, a simulation example demonstrates how the autonomous system determines possible future actions, evaluating them with regard to hazards in order to provide a plan with acceptable risk. Nevertheless, the approach can also be implemented to real world applications since typical real world phenomena as uncertainty and faults can also be considered in the chosen virtual world figuratively.

In this contribution, a novel learning method realizing the refinement of a Cognitive Technical System's pattern recognition and attention capabilities is presented. The method is implemented within a cognitive architecture with a representational level based on Situation-Operator-Modeling and high-level Petri Nets. Through the representational level, it is possible to realize a mental model mapping the complex structure of the real world internally in a compact format reduced to the relevant aspects. The mental model can be created and modified automatically by learning from interaction. If the perceived real world does not correspond to the system's mental model, the system detects ambiguities (or conflicts) inevitably. Then, the system tries to solve the conflicts by a more detailed view to the measured sensor inputs. Thus, new significant features (on a high abstraction level) can be derived from the measurements and taken into account to distinguish different (before apparently equal) situations. The contribution describes the proposed method and its fundamentals in detail. Furthermore, the realization of a cognitive mobile robot is presented as an application example illustrating the proposed method.

In this contribution, a concept and principal realization of an additional module within a proposed HMI analysis architecture is developed. The main aspect of this module is the reduction of complexity allowing the analysis of a Human-Machine-System. Core of the architecture is an action model, which is methodical founded on Situation-Operator-Modeling. The action model describes the interaction within a Human-Machine-System and is implemented by high-level Petri Nets. From the Petri-Net-model a state space can be generated to analyze the interaction between a human operator and the environment (in general assumed as machine). The definition of the situation representing the considered part of the real world influences the size of the state space significantly. This contribution realizes the implementation of a size-variable situation vector to reduce the complexity of the considered system. The functionality of the extended architecture is illustrated by the interaction of a human operator with an arcade game.

A novel architecture for Cognitive Technical Systems with a homogeneous knowledge representation for all cognitive functions is presented. The approach is methodically based on Situation-Operator-Modeling and implemented with high-level Petri Nets. It is not restricted to certain application fields and can be combined with other AI methods. By the combination of several instances of the architecture, the complexity of the represented interaction can be reduced by a problem-oriented encapsulation of different action spaces. The contribution is focused on learning of interaction in and with the outside world and its inner structure, which is illustrated by the control of an arcade game.

The paper presents a concept for the automated supervision of human interactions in environments with complex system dynamics. The supervision concept is formulated using a special system-theoretic Situation-Operator-Modeling (SOM) approach which in the context of this work is practically implemented and simulated using Higher Petri Net (HPN) formalisms and tools. In order to reflect the complex system nature suitable HPN patterns are developed which make use of active tokens with autonomous, continuous-time dynamics. The approach is demonstrated on the supervision of a lane change maneuver in highway driving with special focus on the checking of driver actions with respect to goal-conformance.

An increasing number of tasks within the distributed air traffic management (ATM) system are currently being redesigned towards more cooperative and interactive planning processes involving decision of several agents (automated and human). In order to ensure that agents work together in an optimal manner it has to be achieved that the goals of the individual actors become properly aligned with the global system goal. That way the cooperative system seeks the optimization of the global objective through actors that seek the optimization of their individual self-objectives. Following this approach the paper introduces a game-theoretic perspective to the design of mechanisms for the guidance of arrival traffic at airports. Relevant incentive structures for individual actors are investigated on the basis of results from a distributed air-ground simulation model.

This contribution summarizes the theoretical background of a cognitive-oriented architecture to build autonomous systems. A special situation-operator-model, developed to model the human-machine-interaction, is used to structure the reality and map this structuring to a mental model of the system to enable planning and learning. The presented architecture builds the framework for autonomous behavior, where the mental model of the system is maintained and refined by the cognitive functions. The learning capabilities of the system and the steps to realize the proposed architecture in general are illustrated.

This contribution demonstrates the realization of a cognitive-oriented architecture to build autonomous systems. A special situation-operator-model, developed to model the human-machine-interaction, is used to structure the reality and map this structuring to a mental model of the system to enable planning and learning. The presented architecture builds the framework for autonomous behavior, where the mental model of the system is maintained and refined by the cognitive functions. The steps to realize the proposed architecture are illustrated and detailed for a mobile robot as an example of a cognitive technical system. The application of this modeling technique to autonomously build and update a mental model of the interaction with the environment is demonstrated in detail.

This contribution proposes an architecture of an autonomous system based on a situation-operator-modeling technique to realize a cognitive-based control. The cognitive functions of the system are divided into the learning, testing, and exploration modules; and the planning and plan supervision modules. The proposed architecture of an autonomous learning system consists of different parts, e.g. filter modules, cognitive functions, knowledge base, and goal translation. The core of the approach is the representation of changes of the outside-world of the systems as a sequence of scenes and actions, which are modeled using a special modeling technique. The steps to realize the proposed architecture are illustrated and detailed for a mobile robot as an example.

The modeling of human-machine-interaction (HMI) may help to transfer understanding of human control and to apply the developed modeling technique to autonomous technical systems such as mobile robots. The task of this new kind of intelligent control is to respond autonomously to complex situations primarily specified in words. The paper describes the modeling technique. When modeling HMI AI-like terms are used such as situation, operator etc. to define the system to be considered. The modeling methodology is suitable for modeling human acting, planning, learning and also describing human errors and is able to design an advanced autonomous system.