Evren Daglarli

Behavioral task processing for cognitive robots using artificial emotions

This paper presents an artificial emotional-cognitive system-based autonomous robot control architecture for a four-wheel driven and four-wheel steered mobile robot. Discrete stochastic state-space mathematical model is considered for behavioral and emotional transition processes of the autonomous mobile robot in the dynamic realistic environment. The term of cognitive mechanism system which is composed from rule base and reinforcement self-learning algorithm explain all of the deliberative events such as learning, reasoning and memory (rule spaces) of the autonomous mobile robot. The artificial cognitive model of autonomous robot control architecture has a dynamic associative memory including behavioral transition rules which are able to be learned for achieving multi-objective robot tasks. Motivation module of architecture has been considered as behavioral gain effect generator for achieving multi-objective robot tasks. According to emotional and behavioral state transition probabilities, artificial emotions determine sequences of behaviors for long-term action planning. Also reinforcement self-learning and reasoning ability of artificial cognitive model and motivational gain effects of proposed architecture can be observed on the executing behavioral sequences during simulation. The posture and speed of the robot and the configurations, speeds and torques of the wheels and all deliberative and cognitive events can be observed from the simulation plant and virtual reality viewer. This study constitutes basis for the multi-goal robot tasks and artificial emotions and cognitive mechanism-based behavior generation experiments on a real mobile robot.

Improving human-robot interaction based on joint attention

The current study proposes a novel cognitive architecture for a computational model of the limbic system, inspired by human brain activity, which improves interactions between a humanoid robot and preschool children using joint attention during turn-taking gameplay. Using human-robot interaction (HRI), this framework may be useful for ameliorating problems related to attracting and maintaining attention levels of children suffering from attention deficit hyperactivity disorder (ADHD). In the proposed framework, computational models including the amygdala, hypothalamus, hippocampus, and basal ganglia are used to simulate a range of cognitive processes such as emotional responses, episodic memory formation, and selection of appropriate behavioral responses. In the currently proposed model limbic system, we applied reinforcement and unsupervised learning-based adaptation processes to a dynamic neural field model, resulting in a system that was capable of observing and controlling the physical and cognitive processes of a humanoid robot. Several interaction scenarios were tested to evaluate the performance of the model. Finally, we compared the results of our methodology with a neural mass model.

Artificial Behavioral System By Sensor-Motor Mapping Strategy For Multi-Objective Robot Tasks

In this study, behavioral system based robot control architecture is built up for a four-wheel driven and four-wheel steered mobile robot. Behavioral system is determined as evolutionary neural-fuzzy inference system for behavior generation and self-learning processes in the general robot control architecture. The kinematics and dynamic model of the mobile robot with non-holonomic constraints is used as present structure which is modeled in previous studies. The posture and speed of the robot and the configurations, speeds and torques of the wheels can be observed from the simulation plant and virtual reality viewer. The behaviors are investigated regarding their gains, fuzzy inference structures, real-time applicability and their coordination.

Realtime object detection in IoT (Internet of Things) devices

IoT (Internet of Things) is acommunication network that connects physical or things to each other or with a group all together. The use is widely popular nowadays and its usage has expanded into interesting subjects. Especially, it is getting more popular to research in cross subjects such as mixing smart systems with computer sciences and engineering applications together. Object detection is one of these subjects. Realtime object detection is one of the foremost interesting subjects because of its compute costs. Gaps in methodology, unknown concepts and insufficiency in mathematical modeling makes it harder for designing these computing algorithms. Algortihms in these applications can be developed with in machine learning and/or numerical methods that are available in scientific literature. These operations are possible only if communication of objects within theirselves in physical space and awareness of the objects nearby. Artificial Neural Networks may help in these studies. In this study, yolo algorithm which is seen as a key element for real-time object detection in IoT is researched. It is realized and shown in results that optimization of computing and analyzation of system aside this research which takes Yolo algorithm as a foundation point [10]. As a result, it is seen that our model approach has an interesting potential and novelty.

Rehabilitation applications using brain inspired cognitive architecture for humanoid robots

In this paper, interaction between humans and robots and for rehabilitation applications in social areas is investigated. People suffering from some disorders require better nursing service for interacting socially. Additionally, utilizing human-robot interaction (HRI), this proposed architecture may be a suitable resolution for problems related to optimizing the focusing and sustaining attention states of children with Attention Deficit Hyperactive Disorder (ADHD) and Autism Spectrum Disorder (ASD). In the near future, it is expected that humanoid robots will have more interactive skills in social regions.

Personality identification by deep learning

In recent years, that researchers in psycho-social fields classify the personalities according to different criterias, is one of the most interesting studies. In the viewpoint of the artificial intelligence researches, it is considered that analyzing the personalities will provide achieving realistic character modellings and realizing more intelligent systems via engineering disiplines in future. Beside of methodological gaps and conceptual uncertainities, insufficiencies in the mathematical modellings make developing computational algorithm difficult for this issue. These algorithms can be developed by present numerical or machine learning based methods in the literature. It can be realized by a hybrid method as composition of them. Numerical methods with linear or nonlinear system can also be suitable. From the standpoint of uncertainities, probabilistic (Bayesian, Monte Carlo, etc.) or fuzzy approaches can elaborate the modelling. Machine learning based methods (Markovian, support vector machines, Boltzman machines or artificial neural networks, etc.) can provide benefit to this kind of the study. In this study, we propose deep neural network based personality identification system with dataset which is composed from given responses to a questionaire prepared as suitable to the purpose. Our approach is verified with the classification results related to this.

High altitude smart monitoring system integration by using a helium powered mechanical balloon

Weather Balloons are used excessively in numerous areas and applications today. High altitude balloons are usually unmanned balloons that may climb up to 40km. We propose a high-altitude balloon system design that is capable of self-tuning itself in order to stay at a predefined height limit [1]. This type of a balloon system may be very useful from monitoring geophysical and atmospheric events but also for a vessel to use technological devices such as relay points. This altitude balancing design also allows to be manipulated using a controlled mechanism and may be easily applied for scientific, engineering and industrial purposes. They are much more economic and they almost use no power when they are compared with the alternative technologies.

Design of unmanned semi-autonomous smart probe for near earth space research operations

To conduct studies related to our planet and its atmosphere at high altitudes above 100 km from the ground requires to overcome great challenges. These studies can include meteorological, geographical and astrophysical missions. In this paper, we designed a high altitude unmanned semi-autonomous probe with helium gas balloon for general purpose near Earth space studies. Also hardware architecture of our proposed system includes high performance computing module, Low-level control systems, propulsion/steering mechanisms, radio telemetry system, real-time measurement and data acquisition system, fail-safe mechanisms and real-time vision system. 3D solid design of the system contains helium gas balloon, main flight capsule (craft), peripheral equipment (e.g. cameras, battery packs, radio antenna, etc.) to be connected to the craft and payload capsule. In the future, our probe can be customized according to the other specified missions.

Rehabilitation based computational models for humanoid robots

In this study, communication between humoid robots and humans for human rehabilitation in social environments is researched. People with disabilities such as patients, elderly, babies and infants need intense care in order to communicate socially. More specifically, using human-robot interaction (HRI), this framework can become a suitable solution for problems related to attracting and maintaining the attention levels of children suffering from Attention Deficit Hyperactive Disorder (ADHD) and Autism Spectrum Disorder (ASD). Today, humoid robots find more communicative possibilities in social areas (such as schools, houses, hospitals and care homes) because of their structural compatibility.

Behavior generation strategy of artificial behavioral system by self-learning paradigm for autonomous robot tasks

In this study, behavior generation and self-learning paradigms are investigated for the real-time applications of multi-goal mobile robot tasks. The method is capable to generate new behaviors and it combines them in order to achieve multi goal tasks. The proposed method is composed from three layers: Behavior Generating Module, Coordination Level and Emotion -Motivation Level. Last two levels use Hidden Markov models to manage dynamical structure of behaviors. The kinematics and dynamic model of the mobile robot with non-holonomic constraints are considered in the behavior based control architecture. The proposed method is tested on a four-wheel driven and four-wheel steered mobile robot with constraints in simulation environment and results are obtained successfully.