Συντάχθηκε 16-09-2016 13:49
από Esthir Gelasaki
Email συντάκτη: egelasaki<στο>tuc.gr
Ενημερώθηκε:
-
Ιδιότητα: υπάλληλος.
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ
Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών
Πρόγραμμα Προπτυχιακών Σπουδών
ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ
ΑΛΕΞΑΝΔΡΟΥ ΤΑΝΖΑΝΑΚΗ
με θέμα
Ενεργειακά Αποδοτικός Έλεγχος Βαδίσματος Δίποδων Ρομπότ σε Δυναμικά Περιβάλλοντα
Energy Efficient Control of Bipedal Robot Locomotion in Dynamic Environments
Εξεταστική Επιτροπή
Καθηγητής Μιχάλης Ζερβάκης (επιβλέπων)
Καθηγητής Γεώργιος Σταυρακάκης
Δρ. Νικόλαος Μπεκιάρης-Λυμπέρης
Abstract
Bipedal Robot Locomotion is considered as one of the toughest problems in Control Engineering. In particular, one of the most challenging problems today is the design of extremely energy efficient trajectories, which reflect exactly the constraints of each individual walking step of a biped robot, combined with an effective trajectory tracking control scheme. More specifically, the problem of design and effective tracking control of trajectories that obey the rules of human walking and have the energy efficiency of human locomotion - which is considered the most energy efficient type of walking - is in the spotlight of the robotics and control community particularly for the last year. Unfortunately, no major efforts have been made.
For the abovementioned reasons, in the current thesis, we propose a novel control system, called Decrease & Conquer Feedback Control. The proposed system consists of the Gait Generation and Trajectory Tracking Control Modules.
The Gait Generation Module consists of two Phases. The First Phase calculates walking trajectories for a simplified biped robot (a robot model with fewer degrees of freedom), and the Second Phase (called Energy Efficient Trajectory Synthesis and Verification) based on the calculated trajectories of the First Phase proceeds with the calculation of energy optimal trajectories for the complete biped robot. We implement two variants of the Gait Generation Module, based on two discretization methods: Direct Collocation and Discrete Mechanics.
The Trajectory Tracking Control Module deals with the effective tracking control of the energy-optimal trajectories, under missing velocity signals and disturbances, leading to minimal energy consumption.
The walking capabilities of the biped robot are evaluated through numerous experiments in a variety of terrains, including flat ground, downward and upward slopes, as well as walking downstairs and upstairs. Furthermore, we proceed with the experimental study of the energetics of the resulted bipedal robot walking.
The proposed system, turns out to be extremely effective. It provides the ability to synthesize energy efficient trajectories for the biped-with energy requirements comparable to human walking-, while it applies effective trajectory tracking control under the situation of missing velocities and disturbances, leading to minimal energy consumption. In comparison with the related One Phase methods and high performance commercial solutions, the proposed system is proved to be a reliable, energy efficient and effective way for human-like biped robot locomotion in a variety of terrains and situations.
Τόπος: Λ - Κτίριο Επιστημών/ΗΜΜΥ, 137Π-39,-38
Έναρξη: 19/09/2016 11:30
Λήξη: 19/09/2016 12:30
