In recent advancements in robotics, Artificial Intelligence (AI) methods such as Deep Learning, Deep Reinforcement Learning (DRL), Transformers, and Large Language Models (LLMs) have significantly enhanced robotic capabilities. Key AI models driving advancements in robotic vision include Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), the DEtection Transformers (DETR), the YOLO family of algorithms, segmentation techniques, and 3D vision technologies. Deep Reinforcement Learning (DRL), an AI technique where agents learn optimal behaviors through trial and error interactions with their environment, enables robots to perform complex tasks autonomously. Transformers, originally developed for natural language processing, have been adapted to robotics for tasks involving sequence prediction and data understanding, improving perception and decision-making processes. LLMs leverage vast amounts of text data to enhance robot-human interaction, enabling robots to understand and generate human-like language, thus improving their communicative and collaborative abilities in various applications. The integration of these AI methods enhances the adaptability, efficiency, and overall performance of robotic systems, paving the way for more sophisticated and intelligent autonomous agents.

: Variants of proportional navigation (PN) are perhaps mostly used guidance laws for tactical homing missiles. PN aims to generate commanding missile lateral acceleration proportional to line of sight (LOS) angular rate, so that missile velocity vector rotates in such a way to assure interception of a target. In order to generate commanding lateral accelerations, the guidance system needs measurements of LOS angular rate and the closing velocity between the missile and the target, or the missile velocity. A device which provides guidance information is referred to as the missile seeker. In the case of imaging based seekers (visible light (EO), infrared light (IIR)), LOS rate is estimated using imaging sensor, while closing or missile velocity is measured using appropriate sensors or guess estimated. In this paper, we present the design and simulation of a missile homing system which includes: true PN guidance law, linear multiloop acceleration autopilot, and gimbaled imaging based missile seeker. Target seeker uses advanced deep machine learning object detection YOLO (You only look once) model, for target detection and tracking as well as LOS rate estimation. Comprehensive simulation model, consisting of full 6DOF missile and controls dynamics, 3D world and camera model, is developed. Intensive simulation results show performances of the proposed missile homing system.

Most modern missiles implement some variant of proportional navigation (PN) guidance law. In order to implement this form of navigation, the missile has to measure line of sight (LOS) rate. Devices capable to measure LOS rate are referred to as the seekers. This article aims to present analysis of a missile seeker mathematical models with purpose to obtain LOS rate estimation used for implementation of PN in three dimensions. This paper includes MATLAB simulations of developed seeker with 6-DOF nonlinear missile mathematical model and autopilot presented in earlier works by the authors.

This article presents proportional navigation(PN) and its few variants used in modern tactical missile guidance. This article develops 6-DOF mathematical model and an autopilot for PN guided missile. Full Simulink simulation and animation of PN navigation in three dimensions is shown and discussed.

In this work we adapt and evaluate different solutions for automatic speech recognition (ASR) to be used as an HMI for the assistant robot. Two on-device solutions: Kaldi (DNN-HMM) and Mozilla's DeepSpeech (end-to-end), and three internet service APIs: IBM Watson, Microsoft Azure and Google Speech to Text are evaluated. The systems are adapted to the domain of robot commands and evaluated on a set of expected inputs. As the goal is to retain the ability to recognise general language, the systems are also evaluated on out of domain data.

This work studies hardware and software aspects concerning the implementation and the performance evaluation of a fast real-time embedded Model Predictive Control (MPC). This study is carried out using a low cost and a low power off-the-shelf high performance embedded computing platform (ARM Cortex M) and an off-the-shelf generic embedded MPC software implementation, with a general purpose QP (quadratic programming) solver (MATLAB). As an example, two oscillating masses system subject to input and output constraints is used. Throughout MPC evaluation is done using a hardware-in-the-loop (HIL) experiments, achieving a couple of milliseconds sampling rates.

An accurate and efficient computer-aided mammography diagnosis system plays an important role as a second opinion to assist radiologists. Finding an accurate and robust computer-aided diagnosis system for classification of the abnormalities in the mammograms as malignant or benign still remains a challenge in the digital mammography. In this paper, a fully autonomous classification system is presented and it consists of the three stages. The input Regions of Interest (ROIs) are obtained using an efficient Otsu's N thresholding and further subjected to a number of preprocessing stages. After preprocessing stage, from the ROIs, a group of 32 Zernike moments with different orders and iterations have been extracted. These moments have been applied to the neural network classifier. The experimental results show that the proposed algorithm is efficient comparing to the ground truth table given in the Mammography Image Analysis Society (MIAS) database.

Image processing plays an important role in medical image analysis. The most popular methods for image processing and analysis are very resource hungry, which leads to some disadvantages in their applications even on a powerful desktop computers. On the other side, modern mobile devices are equipped with powerful processors with an efficient instruction architecture. This lead to better performance per watt than a desktop CPUs. This work investigates the performance of a widely used medical analysis algorithm implemented on a modern mobile devices and desktop CPU. The results obtained with ARM NEON instructions show speed improvements up to 2 times. As this research shows mobile devices cannot yet compete with powerful desktop CPUs, even with using highly optimizations or multiple threads. In the last part of the paper conclusions are drawn for acceptable image input and parameter sizes.