Permanent URI for this collection
Browse
Recent Submissions
Item Development of a human-machine interface for teleoperation and a compensation controller for time-delay systems.(Tshwane University of Technology, 2023-07-12) Nahri, Syeda Nadiah Fatima; Prof Shengzhi Du; Prof Barend Jacobus Van WykDeveloping autonomous robots that can assist humans in their daily chores has been among the essential visions in robotic research. Robots must accommodate multi-modal (sensory) to interact with humans effectively. Humans commonly utilize auditory, visual, olfactory, and haptic sensations. Among these, haptic sensations attract the most attention in human-robot bilateral collaboration systems. The environmental information from the slave side (remote) is fed back and displayed through the haptic interface device on the master side, then experienced by the human operator. Using the haptic sensation, a human operator can feel like he/she is touching an object in a remote environment without being physically present. These haptic feedback control systems are typical human-in-the-loop (HITL) systems, a relatively young field of research with certain challenges under study. The controller design and theoretical analysis of such systems are challenging tasks, as a human is a highly uncertain object when involved in human-robot bilateral collaboration systems. Each human has his own experiences of the environment, individual executing movements, and different timings. The time delay of such a system is the main challenge, primarily when the public communication networks (internet) are used during haptic-feedback bilateral teleoperation, along with the challenges from external disturbances and noise present in the system/object to be controlled under time-delay scenarios. These factors significantly affect the transparency and stability of the complete teleoperation system. Hence, when designing a controller for these systems, it is necessary to obtain a reasonable trade-off between the two contradictory goals: the stability and transparency. The purpose of this research is to alleviate the challenges related to time-delay and disturbance compensation; hence this thesis takes the following steps: firstly, a comprehensive review of the various model-based and model-free (also known as model independent) control methods for bilateral teleoperation systems is conducted by highlighting their respective outlines, viability, and limitations. Further, this review gives way to the model-independent Active Disturbance Rejection Control (ADRC) and its time delay and disturbance compensation methods developed over time. This review identifies state-of-the-art methods towards a suitable control system. Secondly, an electro-mechanical platform, i.e., the human-machine interface (HMI) platform prototype, is developed. This platform serves as an interfacing device for the master (human operator) in haptic-feedback bilateral teleoperation systems. Thirdly, the HMI prototype developed is controlled using the traditional Proportional Integral (PI) controller, followed by the model-independent ADRC controller. The results suggest a high disturbance compensation tolerance and robust behaviour of the ADRC compared to the conventional PI control. The scientific contribution of this research is a novel predictive extended observer (ESO)-based ADRC proposed for disturbances and measurement noise compensation in systems with time delay. This is a novel architecture that adheres to the predictive principle for both time delay and disturbance compensation. The proposed predictive extended state observer (ESO)-based ADRC is validated on several case studies and compared with existing Extended State Predictor Observer (ESPO)-based and the modified time-delay-based ADRC controller methods. The experiment results show the efficacy of the developed controller in solving the problems of external disturbance compensation along with measurement noise in the presence of time delay. Moreover, the results on transient response characteristics indicating performance parameters (transparency) were analyzed, displaying an improvement compared to the state-of-the-art methods (ESPO-based and modified time-delay-based ADRC controllers). Different outcomes were achieved with the proposed controller; the results show a decrease in response overshoot by 1 % to 9.1 %, and zero steady-state error was obtained after output disturbance compensation compared to the delay-based ADRC and ESPO-based controller methods, respectively. Further, the stability analysis of the proposed controller was conducted using Bode plots showing positive gain margins and phase margins (one being 30.8 dB and 146 deg, respectively, for the Type 0 system experiment). Furthermore, the phase crossover frequency (38.4 rad/sec) exceeded the gain crossover frequency (1.58 rad/sec), thus proving its stability. Moreover, Nyquist diagrams confirmed the stability of the time delay systems. Future study is recommended to enhance experimental results on time-delay compensation for greater delay values for the suggested predictive controller architecture.