Jakub Mozaryn

Robust velocity estimation for legged robot using on-board sensors data fusion

Availability of momentary velocity of a legged robot is essential for its efficient control. However, estimation of the velocity is difficult, because the robot does not need to touch the ground all the time or its feet may twist. In this paper we introduce a method for velocity estimation in a legged robot that combines kinematic model of the supporting leg, readouts from an inertial sensor, and Kalman filter. The method alleviates all the above mentioned difficulties.

Robust estimation of walking robots velocity and tilt using proprioceptive sensors data fusion

Availability of the instantaneous velocity of a legged robot is usually required for its efficient control. However, estimation of velocity only on the basis of robot kinematics has a significant drawback: the robot is not in touch with the ground all the time, or its feet may twist. In this paper we introduce a method for velocity and tilt estimation in a walking robot. This method combines a kinematic model of the supporting leg and readouts from an inertial sensor. It can be used in any terrain, regardless of the robots body design or the control strategy applied, and it is robust in regard to foot twist. It is also immune to limited foot slide and temporary lack of foot contact. A method of velocity and tilt estimation in mobile, possibly legged robots based on on-board sensors.Robustness to inertial sensor biases, and observations of low quality or temporal unavailability.A simple framework for modeling of legged robot kinematics with foot twist taken into account.

Relative Error Indices for Comparison of Neural Models of Different Robots

In the article there are proposed relative error indices that can be used to evaluate neural models of different robots. The model of robot is identified using artificial neural networks with structure of the mathematical robot model in a form of the Lagrange-Euler equation. As an example the proposed indices are calculated for neural models of three different robots. Proposed indices significantly simplify an analysis and comparison of models of robots with different degree of freedom.

Experimental Evaluation of Mathematical and Artificial Neural Network Modeling of Energy Storage System

This article presents an experimental evaluation based on a mathematical model and an artificial neural network (ANN) model of an energy storage system. Because of a nonlinear description of charging/discharging dynamics in subsequent cycles and a coupling of the terminal voltage and temperatures of a battery, the recurrent artificial neural network structure (R-ANN) is proposed. Both models, analytical and R-ANN were employed to predict a behavior of the VRLA AGM battery. A training and testing data were gathered at the laboratory stand in different working conditions. As a result, we present the analysis of differences between proposed modeling approaches.

Design of a neural network for an identification of a robot model with a positive definite inertia matrix

This article presents a method of designing the neural network for the identification of the robot model in a form of Lagrange-Euler equations. It allows to identify the positive definite inertia matrix. A proposed design of a neural network structure is based on the Cholesky decomposition.

Design process and experimental verification of the quadruped robot wave gait

In this paper there is presented the design process and experimental verification of the quadruped robot wave gait. Mathematical model of a robot movement is a result of linking together derived leg movement equations with a scheme of their locomotion. The gait is designed and analysed based on two-step design procedure which consists of simulations using MSC Adams and Matlab environments and experimental verification using real quadruped robot.

Calculation of Model of the Robot by Neural Network with Robot Joint Distinction

There is presented the design of the feedforward neural network for calculation of coefficients of the robot model. Proposed method distinguishes the degrees of freedom and improves the performance of the network using information about the control signals. A numerical example for calculation of the neural network model of Puma 560 robot is presented.

Modelling of Ultracapacitors Using Recurrent Artificial Neural Network

This article presents an artificial neural network (ANN) model of the ultracapacitors based on experimental data acquired from laboratory purposely built test stand for dynamic load cycle tests. Because of a nonlinear description of discharging dynamics in subsequent cycles and a coupling of the terminal voltage and temperatures of a ultracapacitor, the recurrent artificial neural network structure (R-ANN) structure is proposed. As a result, it was presented the accuracy analysis based on the statistical quality indices of proposed modeling approach.

Model-Based Research on Ultracapacitors.

The following article presents the analytical basis and research results for ultracapacitors, acquired purposely built test stand for dynamic load cycle tests. Furthermore, three methods allowing for determination of ultracapacitor model parameters were presented in the work. First was the off-line identification in time domain method, involving minimization of the mean square error of model response and element response alignment for given input. Second method was off-line identification in frequency domain, involving application of least squares method with implemented model response correction procedure. Third method was on-line identification which allowed for determination of model parameters in real time with use of Kalman filter. The methods presented can be applied in research of other types of energy storage systems, i.e.: electrochemical batteries or hybrid energy storage systems.

Theoretical and experimental background for artificial neural network modeling of alpha type Stirling engine

This article presents a theoretical background for an artificial neural network (ANN) model of the alpha type Stirling engine where thermodynamic dependencies, connected with equations of motion for the piston-crankshaft system with three degrees of freedom were taken into account. Because of the highly nonlinear description of Stirling engine dynamics, the ANN was employed, that modelled output power of Stirling engine as a function of the input power, molar mass, load current, pressure obtained by gas combustion and working parameters of the engine. The ANN model was tested on experimental data, gathered at the laboratory stand, in different working conditions. The proposed ANN model provides good results for both training and testing data-sets.