Approaching Dynamic Gait with AAU-BOT1
Authors
Madsen, Rune ; Andersen, Niels Hyltoft
Term
4. term
Education
Publication year
2013
Submitted on
2013-06-06
Pages
186
Abstract
Dette speciale undersøger, hvordan AAU-BOT1 kan opnå dynamisk gang. Tidligere arbejde har skabt en funktionel robot, hvor hvert led kan styres individuelt; her er fokus at designe og implementere en kontrolstrategi, der får robotten til at gå. For at informere designet blev menneskelig gang undersøgt, herunder optagelser i et bevægelsessporingslaboratorium. På den baggrund blev der afledt kinematiske (bevægelser) og dynamiske (kræfter) modeller af robotten for at bestemme de nødvendige ledbevægelser. Kontrolsystemet har to niveauer. Lavniveau-kontrollere regulerer hvert enkelt led. En høj-niveaus gang- og balancekontrol genererer ledreferencer og holder robotten stabil. Balancen vurderes ved at måle overkroppens vinkel og samspillet mellem fod og underlag. Da fod–underlag-stød kan være skadelige, anvendes en støddæmpningsstrategi til at blødgøre hårde kontakter. Med denne kontrolstrategi var AAU-BOT1 i stand til at gå med en hastighed op til 0,15 km/t.
This thesis explores how to achieve dynamic walking with the AAU-BOT1 robot. Earlier work produced a functional robot with individually controllable joints; the focus here is to design and implement a control strategy that makes the robot walk. To guide the design, human walking was studied, including recording motion in a tracking laboratory. Based on this, kinematic (motion) and dynamic (forces) models of the robot were derived to determine the joint movements needed for walking. The control system has two levels. Low-level controllers regulate each joint. A high-level gait and balance controller generates joint targets and keeps the robot stable. Balance is estimated by measuring the torso angle and the interaction between foot and ground. Because foot–ground impacts can be damaging, a shock-absorption strategy is applied to soften hard contacts. With this controller strategy, AAU-BOT1 was able to walk at speeds up to 0.15 km/h.
[This abstract was generated with the help of AI]
Documents