About us

Ole Magnus Brastein, Ph.D

Brastein started BEDKO AS in 2011. Since then, the company has offered consulting services in software engineering and electronics. He has 15+ years of experience as a programmer and software architect.

His experience range from the fields of machine learning (ML) and system identification (SID) to software and hardware engineering. He is experienced in software development for both PC and embedded systems, and the use of third party tool-sets.

Combining a solid theoretical understanding of cybernetics, control theory, machine learning and system identification, with hands on experience in software development has proven to be a valuable combination for our customers.

In addition to hands on consulting/work on projects, Brastein is available to hold seminars and short talks within his fields of interest.

Full CV:    English        Norwegian

Previous projects


  1. mTrackHeave estimation for active compensation
    • Colaboration with Scantrol AS
    • Algorithm to estimate heave motion of ships
  2. FEED – Data collection tool for marine science use
    • Highly configurable front end for SQL databases
    • Extensive connectability to large range of different assets, many supported protocols
    • Video demo here
  3. iSYMHMI/SCADA for offshore/ship use
    • Colaboration with Scantrol AS
    • Winch and ship control/monitoring software
  4. Auto-tuningSoftware suite for calibration of winch characteristics
    • Colaboration with Scantrol AS
    • Identifies and optimizes a characteristic for winches
    • Improves speed control of the winch by learning its behavior
  5. Babel – Graphical test tool utility
    • Easily creates test tools for communication and data treatment
    • A range of GUI elements to help present data
    • Easily connects to TCP/UDP/USB/Serial units
    • A selection of protocols (NMEA, Modbus, etc)
  6. Greyhound – Dynamic system models
    • Powerfull toolset for working with dynamic system model calibration
    • Calibrate models using a range of objectives (stochastic and deterministic)
    • Analyse the parameter space to detect parameter identifiability and interdependence issues: Profile Likelihood, Bootstrapping and more.

Robotics and autonomy

    1. Hexapod 1 – Autonomous
      • Ultrasonic sensors (I2C), uC based onboard AI
      • Simple, robust mechanical construction, ideal for teaching SW engineering and AI
    2. Hexapod 2 – Autonomous
    3. Robot arm  – Manipulator
      • 6 DOF robot arm, using simple hobby servos
      • uC based motor control, with PC based end point/motion control
    4. Winch model
      • Mini-winch with electric drive, speed control
      • Cascade PI + P control of speed and position w/ feed-forward
      • Configurable, for educational and experimental use
    5. Bi-ped human form robot
      • Work in progress, 18 DOF w/servos and 3D-printed mechanics
      • On-board raspberry PI controller for gait/locomotion and stability control
    6. 3D simulation of inverted pendulum – MATLAB interface
      • OpenGL used to simulate an inverted pendulum
      • Interconnects to MATLAB for MPC control
    7. Quadruped – 12DOF dynamicaly stable walker
      • Work in progress
      • Raspberry PI onboard control of gait/stability
      • Not staticaly stable, requires dynamic motion control


  1. ModbusIO – IO modules with Modbus TCP
    • A series of IO modules (DI/DO/AI/AO/PWM)
    • ModbusTCP interface

Technologies and fields of interest

Software development:

  1. C# (WinForms, UWP)
  2. C (for embedded uC / RISC)
  3. C++ (incl. MFC for legacy projects)
  4. Python
  6. UML and its use in Object Oriented Analysis and Design

Machine learning

  1. Artificial neural networks
  2. Deep learning
  3. Decision trees
  4. Evolutionary algorithms
  5. Ensemble methods (boosting, bagging, Random Forrest, etc)

Cybernetics and algorithms

  1. Kalman Filters (incl. non-linear variants such as EKF, UKF, EnKF)
  2. Stochastic models and parameter estimation
  3. Model Predictive Control
  4. System identification (PEM, SSID, etc)

Applied mathematics

  1. Numerical optimization
  2. Statistics and probability

Software and tools

  1. Visual Studio (v6.0 – 2017)
  2. MATLAB (incl SID and ML tools)
  3. Python
  4. Eagle PCB layout / CAD
  5. Microchip tools for embedded software
  6. LyX (a LaTeX based editor)

Lectures and presentations

  1. Machine learning workshop, SMART group at USN (23.11.17)
  2. Machine learning for the industry, TEKNA (29.5.18)
  3. What is Artificial intelligence? – Ethical challenges for society, GET Academy Larvik (20.11.18)
  4. Machine Learning for the process industry, YARA Advanced Process Control (06.12.18)
  5. Machine Learning – an introduction, Industry-Science MeetUp (22.05.19)


[1] O. Brastein, B.  Lie, R. Sharma, N.-O. Skeie, Parameter estimation for externally simulated thermal network models. Energy and Buildings, vol. 191, pp. 200 - 210, 2019.

[2] O. Brastein, D. Perera, C. Pfeifer, and N.-O. Skeie, Parameter estimation
for grey-box models of building thermal behaviour, Energy and Buildings,
vol. 169, pp. 58 – 68, 2018.

[3] O. Brastein, R. Olsson, N.-O. Skeie, and T. Lindblad, Human activity recognition by machine learning methods, Norsk Informatikkonferanse, 2017.

[4] O. M. Brastein, Grey-box models for estimation of heating times for buidings, Master thesis, Høyskolen i Sørøst-Norge, 2016.