93C85 Automated systems (robots, etc.) [See also 68T40, 70B15, 70Q05]
Refine
Document Type
- Doctoral Thesis (2)
- Diploma Thesis (1)
Has Fulltext
- yes (3)
Keywords
Faculty / Organisational entity
Industrial robots are vital in automation technology, but their limitations become evident in applications requiring high path accuracy. This research focuses on improving the dynamic path accuracy of industrial robots by integrating additional sensor technology and employing intelligent feed-forward control. Specifically, the inclusion of secondary encoder sensors enables explicit measurement and compensation of robot gear deformations. Three types of model-based feed-forward controllers, namely physics-based, data-based, and hybrid, are developed to effectively counteract dynamic effects.
Firstly, a physics-based feed-forward control method is proposed, explicitly modeling joint deformations, hydraulic weight compensation, and other relevant features. Nonlinear friction parameters are accurately identified using a globally optimized design of experiments. The resulting physics-based model is fully continuously differentiable, facilitating its transformation into a code-optimized flatness-based feed-forward control.
Secondly, a data-based feed-forward control approach is introduced, leveraging a continuous-time neural network. The continuous-time approach demonstrates enhanced model generalization capabilities even with limited data. Furthermore, a time domain normalization method is introduced, significantly improving numerical properties by concurrently normalizing measurement timelines, robot states, and state derivatives. Based on previous work, a method ensuring input-to-state and global-asymptotic stability is presented, employing a Lyapunov function. Model stability is enforced already during training using constrained optimization techniques. Moreover, the data-based methods are evaluated on public benchmarks, extending its applicability beyond the field of robotics.
Both the physics-based and data-based models are combined into a hybrid model. Comparative analysis of the three models reveals that the continuous-time neural network yields the highest model accuracy, while the physics-based model delivers the best safety properties. The effectiveness of all three models is experimentally validated using an industrial robot.
Eine grundlegende Voraussetzung für die Entwicklung von teilautonomer hydraulischen Maschinen ist die automatisierte Bewegung der notwendigen Strukturen. Die Anforderungen an diese Bewegungen sind hoch und aus diesem Grund sind einfache Reglerstrukturen ohne konkrete Streckenkenntnis nicht ausreichend. Die Beschreibung des Streckenkennmodells ist bei hydraulischen Maschinen sehr komplex und zeitaufwendig. Die Modellierung mit theoretischen auf physikalischen Ansätzen beruhenden Methoden ist daher unwirtschaftlich. Aufgrund dessen müssen für die Entwicklung von teilautonomen Maschinen alternative Strategien zur Beschreibung der Dynamik entwickelt werden. Im Rahmen der Diplomarbeit wurde die Machbarkeit von Neuronalen Netzen zur modellbasierten Geschwindigkeitsregelung von hydraulischen Zylindern an einem Bagger untersucht. Dabei wurden unterschiedliche Anregungssignale überprüft und das Regelverhalten des Neuronalen-Reglers auf einem realen Versuchsträger verifiziert. Es hat sich gezeigt, dass sich die datenbasierten Methoden zur Regelung von elektrohydraulischen Baggern eignen. Im Vergleich zu theoretischen Ansätzen konnte eine Steigerung der Regelgüte, bei gleichzeitiger Reduzierung des Arbeitsaufwandes von mehreren Monaten hinzu wenigen Tagen, erreicht werden.
Accurate path tracking control of tractors became a key technology for automation in agriculture. Increasingly sophisticated solutions, however, revealed that accurate path tracking control of implements is at least equally important. Therefore, this work focuses on accurate path tracking control of both tractors and implements. The latter, as a prerequisite for improved control, are equipped with steering actuators like steerable wheels or a steerable drawbar, i.e. the implements are actively steered. This work contributes both new plant models and new control approaches for those kinds of tractor-implement combinations. Plant models comprise dynamic vehicle models accounting for forces and moments causing the vehicle motion as well as simplified kinematic descriptions. All models have been derived in a systematic and automated manner to allow for variants of implements and actuator combinations. Path tracking controller design begins with a comprehensive overview and discussion of existing approaches in related domains. Two new approaches have been proposed combining the systematic setup and tuning of a Linear-Quadratic-Regulator with the simplicity of a static output feedback approximation. The first approach ensures accurate path tracking on slopes and curves by including integral control for a selection of controlled variables. The second approach, instead, ensures this by adding disturbance feedforward control based on side-slip estimation using a non-linear kinematic plant model and an Extended Kalman Filter. For both approaches a feedforward control approach for curved path tracking has been newly derived. In addition, a straightforward extension of control accounting for the implement orientation has been developed. All control approaches have been validated in simulations and experiments carried out with a mid-size tractor and a custom built demonstrator implement.