Swing up Control of an Inverted Pendulum
Authors
Kapnisakis, Ioannis ; Catarinacci, Francesco
Term
4. term
Education
Publication year
2016
Submitted on
2016-10-31
Abstract
Dette speciale omhandler sving-op og stabilisering af et inverteret pendul på en vogn. De elektriske og mekaniske delsystemer analyseres, og der opstilles en matematisk model for både aktuering og den underaktuede dynamik. Med udgangspunkt i pendulets totale energi udvikles en energibaseret sving-op-strategi, som nær den oprette position skifter til en lineær regulator, der stabiliserer omkring den ustabile ligevægt. Lyapunov-baserede overvejelser anvendes i formuleringen af styreloven. Tilgangen vurderes i Matlab-simulationer og implementeres på en laboratorieopstilling med mikrocontroller, servo-forstærker, DC-motor, vogn-pendul og encodere, herunder praktisk tuning af sving-op og skiftekriterier. For at forbedre tilstandsestimationen designes et udvidet Kalman-filter off-line med modellering af måle- og systemstøj, og resultater heraf præsenteres. Specialet omfatter desuden parameterestimering og håndtering af friktion. Simulationer og eksperimentelle resultater præsenteres; den medfølgende uddragsdel angiver dog ikke kvantitative præstationsmål.
This thesis addresses swing-up and stabilization of a cart–inverted pendulum. The electrical and mechanical subsystems are analyzed and a mathematical model is derived for the actuation and underactuated dynamics. An energy-based swing-up strategy that regulates the pendulum’s total energy is developed and, near the upright position, the controller switches to a linear stabilizer around the unstable equilibrium. Lyapunov-based arguments guide the control-law formulation. The approach is evaluated in Matlab simulations and implemented on a laboratory setup comprising a microcontroller, servo amplifier, DC motor, cart–pendulum, and encoders, including practical tuning of the swing-up and switching logic. To improve state estimation, an extended Kalman filter is designed off-line with measurement and process noise modeling, and its results are reported. The thesis also covers parameter estimation and friction effects. Simulation and experimental results are presented; detailed performance figures are not provided in the excerpt.
[This summary has been generated with the help of AI directly from the project (PDF)]
Documents
Other projects by the authors
Kapnisakis, Ioannis:
Catarinacci, Francesco:
