WIRELESS HYBRID VISUAL SERVOING ALGORITHMS FOR MOBILE ROBOTS

Abstract

In this thesis, some novel algorithms for wireless hybrid visual servoing of mobile robots has been designed to control the motion of omnidirectional wheeled mobile robots. In particular, applications like trajectory tracking, searching and tracking of targets to maneuver within static and dynamic obstacles and minimizing tracking error are considered. The proposed algorithms are based on the hybrid visual servoing consisting of position-based visual servoing (PBVS) and image-based visual servoing (IBVS). The controllers associated with these algorithms acquire information about the environment and give an optimum path planning to perform a collision free tracking. Additionally, these controllers recognize whether there exist any obstacle or neighboring robots (for a swarm of mobile robots). In all the approaches, neural extended Kalman lter (NEKF) is implemented to reduce any uncertainty or noises existing in the system. A novel dynamic model and its control with stability analysis for an omnidirectional wheeled mobile robot are considered. In particular, a dynamical model is described which uses a feedback linearization to control the omnidirectional mobile robot in the tracking of a given trajectory. A PID controller is considered for controlling the motion of the mobile robot. In order to prove the stability of the proposed system, the Lyapunov method is used for showing the asymptotic stability. The simulations and experimental results are carried out with the dynamic control method which leads the stability of the mobile robot during circular path navigation. A new control algorithm for wireless hybrid visual servoing of an omnidirectional wheeled mobile robot is proposed. The hybrid system is developed using the two autonomous control algorithms, i.e., PBVS and IBVS. For the linearization of output signals, neural network extended Kalman lter is used. Several experimental as well as simulation results are presented in order to show the applicability of the proposed algorithm. It is always challenging to track a moving object by a mobile robot. In this context, a new optical ow-based hybrid visual servoing algorithm for tracking of a moving object with an omnidirectional wheeled mobile robot is proposed. A depth observer is taken into consideration for the fact that the object is not stationary. The existed disturbance on the dynamics of visual feature due to the object motion has been estimated. Moreover, a compensation term is added in the vision based controller. A number of experimental and simulation results are presented to validate the proposed algorithm. It is interesting to see the applicability of the above control algorithm in tracking of multiple moving objects by a swarm of robots. A new hybrid visual servoing algorithm for a swarm of mobile robots to track multiple targets is proposed. To solve this problem, a distributed method and the combination of IBVS and PBVS to improve the performance has been used. A neural network extended Kalman lter (NEKF) is used for reducing noises existed during the motion of the object. The control method consists of two approaches: interaction locally among robots and target tracking. Infrared proximity sensors and monocular cameras are applied to ful ll these requirements. Simulation results validate the applicability of the proposed algorithm in tracking of multi-target using a swarm of robots.