Abstract
Nowadays, mobile robotic systems use many complex electronic devices that employ a wide range of control and signal processing techniques to capture and process their environment in real time. Mobile robots use various types of sensory parts, such as CCD cameras, laser scanners, infrared, and ultrasonic sensors, as well as IMU and GPS devices. Also, a combination of the aforementioned sensory parts is used, which determines or measures the actual physical parameters of the environment as well as the current global position, orientation, and velocity of the mobile robot. Using mixed measurements of the actual physical parameters, different hardware processing units are utilized, implementing the suitable control algorithms to determine the control signals for the robot actuators and taking the decisions based on external conditions of the robotic system. Mobile robots, therefore, play a crucial role in modern industrial applications, as well as in others research and development applications.
During the recent decade, unlike past robotic history, the word “robot” has gone through impressive mental transformation: In the mind of society and common people, it has gone from the meaning in science fiction and latest aerospace projects into a regular gadget, which can be found on the streets and even in homes. We are referring to the cleaner robots, smart domestic security and homework systems, handicap manipulators with voice control, and so on.
As any other epochal phenomenon, obviously it has both positive and negative components. The positive one is that for common people the image of a rebel robot and dangerous machine is not there anymore. People can see that robots are now useful and sympathetic home inhabitants. However, the extremely important issue is that most of well-known fears of robotic fiction in real life is usually related to practical problems. The first of them is that of humans and robots interacting in the same area, and mutual interaction in dynamic environments. Of course, it is not so dramatic as it is in fiction. Robots do not cause disasters, yet physical body contacts or accidental collisions of robots and humans are not preferred. It causes many inconveniences and reduces the effectiveness when fulfilling their respective tasks simultaneously.
Another great problem is recognition of environment by robots. It has many related aspects, such as erroneous distance interpretation, misunderstanding of object images, or detection faults of dangerous objects/situations among them.
Of course, the robotic behavior, individual as well as in swarms, is a very complex area and has many difficult aspects. However, even the briefest analysis of the two abovementioned groups of problems shows us that they have the greatest priority and immediateness in a queue in order to harmonize the coexistence of robotic and biologic objects in the same space. The individual trajectories of the movement of robots and humans must be calculated in order so that it does not cause any inconveniences and undesired collisions during their performance and mutual assistance.
It is obvious then that the relaxation of both problems could also be approached in two different ways: – The improvement of optical, mechanical, and electronic aspects of the machine vision system design. – The design of novel control algorithms for signal processing, taking into account that the vision system can never be perfect. In order to filter out or decrease the errors of sensory parts for the obtention of regularized data.
The present edition takes into consideration both of these meaningful approaches, and their specialties make up two proportional parts of this text, which can be attractive for a wide auditory of specialists. In order to cover this interest in the most complete and diverse way, many versatile situations and problem components are considered in this Special Collection, as is mentioned below.
Thereby, this special collection deals with contributions of control theory applications and signal processing in machine vision systems for automatic navigation of mobile robotic systems. It is intended to provide a reference on machine vision supporting techniques and three-dimensional (3D) reconstruction for researchers and engineers. These contributions are focused on terrestrial and aerial mobile robotic systems applied for technical vision, intelligent navigation algorithms, industrial manufacturing and inspections, and digital controllers; particularly on applications of unmanned aerial vehicles, drones, autonomous vehicles, and mobile humanoid robots.
Topics included in this special collection are the following: Control theory application and implementation in mobile and aerial robotic systems. Signal processing application and implementation in mobile and aerial robotic systems. Instrumentation and control for mobile robotic systems in science and industrial applications. Measurement of actual physical parameters of the surrounding environment using different kinds of sensors. Implementation of computer algorithms on modern digital systems.
In the following, some key aspects are described that the Special Collection deals with. One focus is defined by the use of Kalman filters to solve the problem of simultaneous localization and mapping of mobile robots. The development of a neuro-fuzzy-based adaptive extended Kalman filter technique is presented, as an adaptive neuro-fuzzy extended Kalman filter to minimize the difference between the actual and theoretical covariance matrices of the innovation consequence is developed. In experiments, mobile robotic navigation with multiple landmarks is evaluated using two benchmark situations.
The Special Collection presents work about the development of microelectromechanical systems, based on magnetohydrodynamic for micro-robot applications. It is shown that the micro-flow channel design and its performance under the influence of the Lorentz force is a critical challenge. Thereby, the improvement of a particle image velocity measuring method in an image processing system is proposed. The accurate measurement of the 2D velocity profile is fundamental owing to the requirement of future enhancement of the customized machine vision system to construct the 3D velocity profile of the magnetohydrodynamic stirrer prototype. The presented methodology can be used to measure and validate the design of microelectromechanical systems and other devices that require micro-flow manipulation.
Another focus of the Special Collection is a real-time detection of passengers in modern traffic flows. Thereby, fundamentals for the development of a system that supports the cognitive performance of drivers to prevent accidents in modern societies with elderly people are surveyed. Recently, the development of devices for visually providing information, such as smart glasses or head up display, is in progress. Therefore, the selection problem of information to be presented for drivers to realize the cognitive support system is discussed in the Special Collection.
Furthermore, the Special Collection presents an online control programming algorithm for human–robot interaction, in which robot actions are controlled by recognizing human gestures. This system allows free programming of repetitive robot tasks in real time. An online robot control experiment based on the visual gesture recognition algorithm is performed. Extensive tests have confirmed the effectiveness and efficiency of the presented method.
Another key aspect, worth mentioning, is defined by the examination of the influence of the phase when measuring frequency signals generated by sensors in embedded systems. The principle of rational approximation is presented and analyzed, which has advantages over other frequency measurement methods and which can be used to estimate the frequency of the sensor output signal.
Last but not least, the Special Collection deals with new biased estimation, to solve ill-posed problems more effectively. Thereby, ill-posed least square problems are problems that arise frequently in many engineering applications. Furthermore, the Neumann series formally represents a geometric series, which is used in functional analysis to solve operator equations. Numerical analysis and results show that the proposed biased estimation method based on the Neumann series has improved accuracy compared to the existing robust estimation methods.
As has been shown, the Special Collection deals with current research topics from various engineering fields. More topics, not mentioned here, from control theory application and implementation, signal processing, robot systems, and so on are dealt with in the Special Collection.