Lubomir Dimitrov

Homography-based face orientation determination from a fixed monocular camera

This paper presents a framework for determining the orientation of human faces with a fixed monocular camera which can be used for the purposes of the gaze tracking afterwards. We use homography relation between two views/frames to handle with the lack of depth information. In order to compensate for the lack of depth information in the relationships between the 2D images in the image plane and the 3D Euclidean space, we present a complete vision-based approach to pose estimation. The homography relates corresponding points captured at two different locations of the face and determines the relationships between the two locations using pixel information and intrinsic parameters of the camera. In order to determine the mapping between the two images, it is assumed that in each frame in the video sequence, we are able to locate, extract and labeled four feature points of the face located at virtual plane attached to the face. Face detection and facial feature extraction are executed with Viola-Jones method. The verification stage for face detection use combined cascade of neural network classifiers uses the convolutional neural network.

The use of bypass channel for feedback control of oscillatory object well-known as difficult one for control

This paper solves the problem of designing of regulator for control of complex unstable object. Calculation was done by numerical optimization using control structure comprising bypass channel. The idea of the bypass channel develops the idea of Smith predictor, but, unlike a predictor, its model can be calculated by the more flexible algorithms. Furthermore, this method does not require the presence of a pure delay link in the object model. This makes the bypass channel more flexible and more efficient than a Smith predictor. The object model is taken from list of well-known object models, which are very difficult for the feedback control due to the strong tendency to fluctuations. Earlier attempts to design regulators for this object are unsatisfactory. This paper shows the efficiency of the bypass channel for the said task. The results are confirmed by simulation.

Adaptive systems based on competitive quality criteria

The paper discusses new principle of the organization of the adaptive system. It is based on the use of two quality criteria, which are changed for the worse as a result of the increment in the adjustable parameters in the different directions. Namely, if the increase of this parameter leads to deterioration of one of the criteria, it must lead to an improvement in another criterion. The combined effect of the two adaptation mechanisms should lead to the establishment of equilibrium or small fluctuations around the equilibrium state of adjustable parameters. If the second criterion is missing or can not be measured, it can be used factor of the time instead of it. For example, if the only criterion tends to reduce the adjustable parameters, then with the passage of some time, gradual or stepwise increase of this parameter should occur. The paper also gives practical example of a system based on this principle.

Application of ultrasonic sensor for measuring distances in robotics

Ultrasonic sensors allow us to equip robots with a means of perceiving surrounding objects, an alternative to technical vision. Humanoid robots, like robots of other types, are, first, equipped with sensory systems similar to the senses of a human. However, this approach is not enough. All possible types and kinds of sensors should be used, including those that are similar to those of other animals and creations (in particular, echolocation in dolphins and bats), as well as sensors that have no analogues in the wild. This paper discusses the main issues that arise when working with the HC-SR04 ultrasound rangefinder based on the STM32VLDISCOVERY evaluation board. The characteristics of similar modules for comparison are given. A subroutine for working with the sensor is given.

Model study of automatic and automated control of hysteretic object

This paper presents the results of research and simulation of feature automated control of a hysteretic object and the difference between automated control and automatic control. The main feature of automatic control is in the fact that the control loop contains human being as a regulator with its limited response speed. The human reaction can be described as integrating link. The hysteretic object characteristic is switching from one state to another. This is followed by a transient process from one to another characteristic. For this reason, it is very difficult to keep the object in a desired state. Automatic operation ensures fast switching of the feedback signal that produces such a mode, which in many ways is similar to the sliding mode. In the sliding mode control signal abruptly switches from maximum to minimum and vice versa. The average value provides the necessary action to the object. Theoretical analysis and simulation show that the use of the maximum value of the control signal is not required. It is sufficient that the switching oscillation amplitude is such that the output signal varies with the movement of the object along both branches with hysteretic characteristics in the fastest cycle. The average output value in this case corresponds to the prescribed value of the control task. With automated control, the human response can be approximately modeled by integrating regulator. In this case the amplitude fluctuation could be excessively high and the frequency could be excessively low. The simulation showed that creating an artificial additional fluctuation in the control signal makes possible to provide a reduction in the amplitude and the resulting increase in the frequency of oscillation near to the prescribed value. This should be evaluated as a way to improve the quality of automated control with the helps of human being. The paper presents some practical examples of the examined method.

Calculation of PID-regulator for MISO system with the method of numerical optimization

Control of the object with one output using many inputs is widely used for combining the advantages of every input and eliminating of disadvantages of two or more control channels. Such systems are called MISO, which means Many Inputs — Single Output. For the calculation of such systems we have proposed a theory based on the logarithmic frequency response. This method has proved to be excellent in laser physics and in many other systems where a single channel for control is insufficient. As a rule one of the channels has higher speed but small range of the output changing value. The other channel allows changing of the output value in a wide range but with a restricted speed. Analytical calculation method cannot be used always as if an object comprises a nonlinear units or delay units there is no analytical methods for this case. Most effective in this case is the numerical optimization method but even when it is used some unexpected problems may occur. This paper studies the regulator design method for such a system. It is shown that the numerical optimization method allows calculation of the required two-channel regulator only with the use of some specific approaches that are kept to a list of recommendations. The results are illustrated by numerical examples obtained by using the program VisSim.

The use of sigma-delta-ADC in the commutation mode

Analog-to-digital converters (ADC) in switching mode are used to convert the signals from multiple sources of the ADC means. It was considered appropriate, since it was believed that this provides a significant simplification of the hardware. Today, the advantage of hardware saving is negligible in compare with the loss of measurement accuracy because of the use of switching mode, which is unreasonably high. ADC developers still offer switching mode, include it in the realization of the circuit performance and recommends it in the user hides. This paper describes carried out comparative study of the ADC in switching mode and this without switching. This study is done with VisSim program simulation. To simulate the operating principle sigma-delta-ADC were used because such ADC are becoming more widespread due to their low cost, high accuracy and speed, sufficient for many practical cases. The simulation results substantiate the reason of non-switched regime. If, however, switching is inevitable, the paper provides recommendations to reduce the error generated by the switching.

Investigation of compensation and other methods for controlling of oscillating objects

The task of control of object, prone to oscillation is still relevant. It has not been solved successfully yet. In particular, it is important to control the object having in the polynomial numerator of its transfer function negative coefficient of the term of the first order [1]. This or close task with varying degrees of success has been solved in a number of papers [1-14]. For example, it is proposed to use a bypass channel, which is a development of the idea of Smith predictor [1, 2]. The use of the bypass channel eliminated the reverse overshoot, but the transient process has preserved a significant reverse deviation and a few oscillations of large amplitude. It is also proposed to use a switching regulator, which at the beginning of the transient process has some initial mathematical model, and changes it to another model at the end of the process or at its middle. In this case, there is an additional transient process at the time of switching, which can be smoothed by mean of the smooth switching. In addition, this method is not effective in the fight with disturbance, because for it the individual stages of the transient process (the beginning, the middle part and the end) can not be separated. The paper investigates alternative methods of control such objects.

Longitudinal Adaptive Control of Mobile Robots in a Platoon

In this paper, In this paper, we present an adaptive longitudinal controller for an autonomous mobile robot platoon, where the only information that the follower robot can use for feedback control is the distance measurements between the robots obtained from a laser range finder. The control velocity of the following robot is computed using the leader velocity estimates obtained from the dynamic (adaptive) part of the controller. Stability analysis of the closed-loop control system is presented. Simulation results of some of the basic vehicle following maneuvers are given to verify the performance of the designed controller. Experiments that demonstrate the performance of the proposed controller are also included.

Some Practical Problems of Distribution and Equalization of Internal Loads in Planetary Gear Trains

The reasons for the uneven distribution of load between the planets in planetary gear trains and the methods for its experimental determination are discussed. An analysis of influencing factors is made up. Below are ways for solving of various problems in the design of one extreme responsible gear train from the mechanism of lifting 200-tons load.