Sahu, Umesh Kumar (2020) Adaptive and Vision Based Controllers for a Flexible Link Manipulator. PhD thesis.
|PDF (Restricted upto 26/02/2023) |
Restricted to Repository staff only
In recent years, Flexible¬Link Manipulators (FLMs) find a wide spectrum of applications including space exploration, defense and medical services owing to several advantages over the rigid manipulators. These advantages are, high payload¬to¬mass ratio, lower actuation, high¬speed operation, more maneuverability and transportability, and reduced power consumption. However, in view of the flexible structure of the links in these manipulators, a number of control complexities arise. Owing to non¬collocated sensors and actuators, FLM behaves as a non¬minimum phase system. In order to represent the dynamics of such FLMs accurately, it is necessary to consider an infinite number of flexible modes in its distributed parameter model. However, to facilitate controller realizability higher order modes are truncated. Model uncertainty due to truncation of flexible modes in its dynamics leads to inaccurate tip¬tracking performance and system stability. Therefore, it is challenging to design a control scheme to achieve perfect tip¬tracking performance with small tracking error. Unlike a rigid robot manipulator, tip¬tracking of FLM is difficult due to distributed link flexure. The tip¬tracking problem of FLM can be divided into two sub¬problems, i.e., (i) tracking of tip position, (ii) suppression of the oscillation in flexible links. To control the tip position and suppress the oscillatory motion, it is necessary to measure tip position accurately. Standard mechanical sensors such as strain gauge, accelerometer, etc. are usually employed for measurement of tip point position. However, direct measurement of tip position by these mechanical sensors has estimation error because of incorrect information of physical parameters. The use of vision sensor can be a better substitute of mechanical sensors because it provides an indirect measurement of tip point deflection. Therefore, the objective of the thesis is to design adaptive and vision based tip¬tracking control strategies for a Two¬Link Flexible Manipulator (TLFM).
To achieve accurate tip¬tracking performance in face of model uncertainties and disturbances, a Sampled¬data Extended State Observer based Backstepping (SD¬ESO based BS) controller is proposed. The convergence and stability of the proposed SD¬ESO based BS controller are investigated by using Lyapunov theory
SD¬ESO gives a promising tip¬tracking performance in face of uncertainties, so, it is also used to estimate the uncertainties arising due to packet dropouts of Network Controlled Two¬link Flexible Manipulator (NC¬TLFM). Further, a Digital Smith Predictor (DSMP) based Backstepping (BS) control scheme has been developed for time¬delay compensation. Convergence and overall stability of the proposed SD¬ESO with DSMP based BS control scheme are investigated using Lyapunov theory. Also, the performance of the proposed controller is compared with conventional Proportional Integral Derivative (PID) and Non-Uniform Predictor¬Observer (NUPO) based controllers using simulation and experimental studies.
Control of FLM using Visual Servoing (VS) is motivating in view of achieving the monitoring of the tip position more precisely. The last decade witnessed a great deal of research interest in visual servoing based control of FLM. The use of camera in flexible manipulator control makes it reliable and faster for a wider range of potential applications.
Among different VS control schemes, Image¬Based Visual Servoing (IBVS) is more effective. However, there are many challenges in IBVS scheme such as singularities in the interaction matrix and local minima in trajectories that affect the system performance in real¬time applications. To resolve these issues, image moment based visual features have been designed for IBVS, and the new two¬time scale IBVS controller based on image moment is developed for tip¬tracking control of TLFM. TLFM dynamics is decomposed into two¬time scale models, namely slow and fast models. A shifted moment based IBVS controller is designed for the “slow” subsystem, and a Linear Quadratic Gaussian (LQG) controller is designed for the “fast” subsystem for tip¬tracking control of TLFM. The effectiveness of the proposed new two¬time scale IBVS controller is evaluated by pursuing numerical simulations.
Although the proposed new two¬time scale IBVS controller based on image moment is developed to address the singularity and local minima issue of IBVS, but sometimes object may leave Field¬of¬View (FOV) that leads to failure of visual servoing task. Therefore, to deal with these issues of IBVS, an Adaptive Intelligent IBVS (AI¬IBVS) controller for TLFM is developed. In the proposed controller, an actor¬critic based off¬policy reinforcement learning controller is developed to keep the object within FOV and shifted moment based IBVS controller to complete the visual servoing task. Simulations have been performed to investigate the performance and robustness of the proposed AI¬IBVS controller.
In the thesis, an adaptive SD¬ESO based BS controller is proposed to handle parameter uncertainty such as un¬model dynamics and disturbance. Further, to improve the rejection of noise that occur due to the usage of the mechanical sensor, a new two¬time scale IBVS controller is proposed. Further, to address the issue of IBVS along¬with retention of object in the FOV, an AI¬IBVS controller is proposed.
The adaptive and vision based controllers proposed in this thesis can be applied to many potential applications such as in hazardous and highly radioactive environments, space exploration, disaster management from a safer distance, and for damping of oscillations for similar vibration systems.
|Item Type:||Thesis (PhD)|
|Uncontrolled Keywords:||Adaptive Control; Flexible Manipulator; Image Moment; Observer Based Control; Reinforcement Learning; Network Control System; Tiptracking Control; Visual Servoing|
|Subjects:||Engineering and Technology > Electrical Engineering > Image Processing|
Engineering and Technology > Electrical Engineering > Image Segmentation
|Divisions:||Engineering and Technology > Department of Electrical Engineering|
|Deposited By:||IR Staff BPCL|
|Deposited On:||26 Feb 2021 12:19|
|Last Modified:||26 Feb 2021 12:19|
|Supervisor(s):||Patra, Dipti and Subudhi, Bidyadhar|
Repository Staff Only: item control page