Mikhail Katliar

Alumni of the Department Human Perception, Cognition and Action
Alumni of the Group Motion Perception and Simulation

Main Focus

PhD student - Motion Perception and Simulation research group

In the  research group we are investigating the process of human self-motion perception and how this knowledge can be exploited in motion simulation.

I am interested in development of real-time Model Predictive Control (MPC) methods and their application to motion simulation. MPC allows better reproduction of vestibular stimuli by the motion platform, which significantly increases realism of simulation compared to the traditional approach based on motion filtering.

Model Predictive Control of Motion Simulators

Motion simulators are widely used for pilot training and vehicle performance analysis. Realism of reproduced motion is limited by the physical constraints of the motion system, such as maximum positions, velocities and accelerations of the mechanical axes. This is especially true when high or sustained accelerations need to be reproduced. The challenge is to select a motion that remains within the simulator capabilities but is as close as possible, according to some defined criteria, to the motion in a real vehicle. Current approaches (based on linear filtering) do not guarantee optimality of the motion.

I develop a simulator control algorithm that would make the motion as realistic as possible (optimal), while staying inside the simulator’s constraints. In the simplest case, it means minimizing the difference in rotational velocities and accelerations a pilot or driver is subjected to between a simulator and a real vehicle (Fig. 1). Once developed, such algorithm can be easily adjusted for perception-based simulation, aiming at minimizing the difference between perceived motion in a simulator and in a vehicle.

Fig 1. Accelerations and rotational velocities in a car (red) and in CyberMotion Simulator (blue) for the car roundabout maneuver. The simulator motion is computed by a large-scale optimization algorithm. The difference between the two motions is minimized while satisfying physical constraints of the simulator. f – specific force, ? – rotational velocity.

Past projects

WABS: Perception-based motion simulation

The WABS project („Wahrnehmungsbasierte Bewegungssimulation“) aimed at developing a novel approach to motion cueing. This approach is based on the idea that motion cueing can be improved by including insights from human self-motion perception research in the motion cueing algorithms. This “perception-based motion cueing” (PBMC) approach should increas the realism and quality of motion simulations by exploiting the limitations and ambiguities of the human perceptual system.

More about 

Within this project, I used direct multiple shooting methods to calculate the optimal simulator motion to reproduce a desired vehicle motion. It was calculated offline based on pre-recorded inertial data, obtained in a real vehicle (a helicopter or a car), or on a synthetic data from the CarSim software. After that, the motion was played in the , and its subjective quality was evaluated using the psychophysical method of magnitude estimation.

Applying optimization methods to motion cueing resulted in better use of simulator capabilities and improved subjective experience compared to the classical filter-based approach. Using knowledge about human perceptual system dynamics promises to provide even better realism compared to purely physical-based approach.

Curriculum Vitae


Oct 2004 – Sep 2005

Postgraduate student at Institute of Physics, Polish Academy of Sciences, Warsaw, Poland; Laser Spectroscopy Laboratory.

Oct 2002 – Jul 2004

Postgraduate student at the United Institute of Informatics Problems, Minsk, Belarus; The Laboratory of Self-Organizing Systems

Sep 1996 – Jun 2001

Belarusian State University, Radiophysics and Electronics Department

Diploma Honours in Radiophysics

Specialized in Statistical Signal Processing

Professional Experience

Jun 2012 – Mar 2013

Senior Software Engineer

“Top Systems”, Minsk, Belarus – Moscow, Russia

Project: Geometric kernel for a new generation CAD system. Surface mesh generation algorithms.

Jul 2010 – Jan 2012

Systems Architect

“Softeq Development”, Minsk, Belarus

Project: Design and development of firmware for SanDisk flash memory products.

Feb 2009 – Apr 2010

Assistant Chief of Software Development Department

“TORA” Corp., Minsk, Belarus

Project: 3D visualization system for Su-25 attack plane training simulator.

Nov 2006 – Jan 2009

Scientific Researcher

“Speech Technology Ltd.”, Minsk, Belarus

Project: Off-line speaker diarisation system.

Oct 2004 – Sep 2005

Research Assistant

Institute of Physics, Polish Academy of Sciences, Warsaw, Poland

Project: Development of statistical methods for estimation of isolated liquid droplet parameters from light-scattering data;

Project: Study of charged body movement in electromagnetic Paul trap.

Oct 2002 – Sep 2004

Research Assistant

The United Institute of Informatics Problems, National Academy of Sciences of Belarus

Project: Research on statistical methods for dynamical systems identification.

Nov 2001 – May 2002

Scientific Developer

“AVNEX Ltd.”, Minsk, Belarus

Project: Software for high-quality real-time voice transformation.

Mar 2000 – Nov 2001

Team Leader

“Sakrament Ltd.”, Minsk, Belarus

Project: Russian speech recognition system.

Skills and Interests

Mathematical modelling

Theory of probability, Bayesian inference, system identification, statistical estimation, digital signal processing.


Professional software developer. Expert in in C++ programming (17 years of experience). Strong knowledge of OOA&D. Proficient in MATLAB and Mathematica. Familiar with OpenSceneGraph, OpenGL, GLSL.

Go to Editor View