Search

Unmanned Aerial Vehicles (UAVs), also known as drones, have had an exponential evolution in recent times due in large part to the development of technologies that enhance the development of these devices. This has resulted in increasingly affordable and better-equipped artifacts, which implies their application in new fields such as agriculture, transport, monitoring, and aerial photography. However, drones have also been used in terrorist acts, privacy violations, and espionage, in addition to involuntary accidents in high-risk zones such as airports. In response to these events, multiple technologies have been introduced to control and monitor the airspace in order to ensure protection in risk areas. This paper is a review of the state of the art of the techniques, methods, and algorithms used in video, radiofrequency, and audio-based applications to detect UAVs and Unmanned Aircraft Systems (UAS). This study can serve as a starting point to develop future drone detection systems with the most convenient technologies that meet certain requirements of optimal scalability, portability, reliability, and availability.

The delay in control systems is a feature frequently in real systems due to the transport of objects or information, a series connection of multiple systems or own processing and sensors delay, among others. Recently there have been several studies to identify the external delay MIMO systems, these works are focused on identification and on-line control of MIMO systems and use a multimodel structure based on modified Smith predictor using different search method. It is clear that for the implementation of the algorithm, and to obtain the convergence and stability analysis, it is necessary to have closed-loop equations of modified Smith predictor. However, in these works is not presented the analytical procedure, not be the main object, displaying only the closed loop equations without the procedure for obtaining it. Therefore, to respond, in this paper, we present an analytical way to derive the closed-loop equations of a modified Smith predictor.

Color is a perceptual characteristic, derived from the light reflected on the object and as such does not represent a physical property. However, progress has been made in building mathematical models to define it more clearly. Light sources play a key role in color perception, and one of the measures to characterize the effect is the Color Rendering Index (CRI). The CRI is a measure used to characterize the color reproduction ability of a light source in particular in comparison with ideal light source. In this work, initially developed an algorithm to determine the CRI from SPD and then used a low-cost RGB sensor to find a model that would estimate the value CRI directly. We used four groups of light sources with technical information available, we measured the SPD and found a transformation matrix to convert RGB to XYZ space. This work is framed in the development of tools to control and lighting measurements for environment control tasks and characterization of artificial sources.

This paper presents a method based on static patterns for robotic navigation using computer vision techniques implemented on CMUcam3 embedded system, also presents an analysis of computational complexity for this embedded system. The pattern to be determined is a black line with intersections, driving the robotic agent over the line. A segmentation algorithm based in threshold is used after the images acquisition by the CMUcam3, then defines the thinning patterns and then the linear Hough transform is applied to determine the lines, the angles and the type of the intersections. Due to the linear Hough transform is a method that requires high processing, in this project, the range of the angles are limited and the accumulation space is normalized. The results showthat the method developed for the navigation is accurate and reliable, because in 87% of the decision segments it could determine correctly the type of intersection and the correction angle. The CMUcam3 embedded system succeeded to process an image between 0.15 and 0.28 seconds, depending of the type of the intersection.

Orthogonal transformations have been very useful in the characterization and signal processing. In particular Hartley Transform allows for time-frequency representations and vice versa. This paper presents an algorithm for calculating the Discrete Hartley Transform in embedded systems with the objective of minimizing the computational load and storage capacity required. It exploits the similarity with the Discrete Fourier Transform to use a fast algorithm reduces the number of trigonometric functions calculated using the rotation factors (Twiddle factors). In general, the implementation can increase the size of the processing window and increase computational speed compared to direct calculation.

One of the biggest problems in the development of autonomous mobile interfaces is linked to the development of tracking system. It is essential to know the position of the mobile to close the control loop on the strategy used in path planning. This paper presents an algorithm that identifies two sets of mobile robots as well as guiding and locating each mobile in the plane. To this end, we propose the use of segmentation by thresholding techniques in conjunction with binary logic operations. The system consists of a peripheral digital camera that observes the entire plane and a set of mobility, which are controlled by a master system. Mobile Robots in its physical construction are two possible directions of movement and rotation on its axis. From the study, was obtained an algorithm that identifies two sets of mobile equipment and obtains the position and orientation. The system is supplemented by a prediction strategy that detects a possible collision. Finally, we get a system that identifies two sets of Mobile Robots on a cartesian plane and manages their movement.

Implementation of intelligent machines requires of efficient classification systems under limited computational resources. Thisstudy introduces a method for estimating the parameters of Radial Basis Function Neural Network (RBF-NN) that can be implemented on a fixed point processor. First, the number of hidden nodes is chosen based on statistics of the mapped data points. A k-means search is then carried out to determine the location of each node. The hidden units mapping corresponds to the Euclidean distance of their centers to each data point, the weights of the output sum are obtained by solving a linear least squares problem. With this procedure, a low computational cost classifier can be readily implemented on a low capacity platform for real time applications.