Mario Sarcinelli Filho

High-level underactuated nonlinear control for rotorcraft machines

This paper proposes a high-level underactuated nonlinear controller capable to guide a RUAV during a 3D flight. First, it presents a dynamic model to represent the dynamics of the aircraft, explicitly showing its underactuated character. Following, a suitable controller based on partial feedback linearization is designed for stabilizing the rotorcraft dynamics. A proof of the stability of the closed-loop control system in the sense of Lyapunov, including modeled disturbances and parametric errors, is also presented, as well as experimental results obtained with a quadrotor, which validate the proposed model and controller.

Human-Machine Interface Based on Electro-Biological Signals for Mobile Vehicles

In this paper, a system to allow the communication between a human being and a robot, through a human-machine interface (HMI), is proposed. Such HMI makes possible to use electro-biological signals, such as electromyogram (EMG), electrooculogram (EOG) and electroencephalogram (EEG) to control devices like an autonomous wheelchair. An electronic board containing an environmental map has also been developed. Thus, the user selects a cell in the map, through some electro-biological signal, which is understood by the system as the place the mobile vehicle should reach. A control system to guide the robot to seek for this goal is also presented, including experimental results

An approach to avoid obstacles in mobile robot navigation: the tangential escape

An approach to guide a mobile robot from an initial position to a goal position avoiding any obstacle in its path, when navigating in a semi-structured environment, is proposed in this paper. Such an approach, hereinafter referred to as tangential escape, consists in changing the current robot orientation through a suitable combination of the values of the angular and linear velocities (the control actions) whenever an obstacle is detected close to it. Then, the robot starts navigating in parallel to the tangent to the obstacle, regarding the point of the obstacle boundary the robot sensing system identifies as the closest one. The stability of the control system designed according this approach is proven, showing that the robot reaches any reachable goal, with or without a prescribed final orientation. Such a control system is programmed onboard a mobile platform whose sensing system is a laser scanner which provides 181 range measurements, for experimental validation. The results obtained are presented and discussed, allowing concluding that the tangential escape approach is able to guide the robot along trajectories that result in a reduction of the traveling time, thus saving batteries and reducing the motor wearing.

Proposal of Brain-Computer Interface architecture to command a robotic wheelchair

This paper presents a Brain-Computer Interface architecture that is being implemented in a robotic wheelchair. The interface uses electroencephalographic signals and works with three mental tasks, which are the imagination of right or left hand movements and generation of words beginning with the same random letter. This research uses a data set to perform a simulation of real-time classification, which is the pseudo-online technique, in order to have a preliminary view of the performance of the proposed BCI architecture. Linear Discriminant Analysis is used to recognize the mental tasks. The feature extraction uses the Power Spectral Density and the choice of EEG channel and frequency uses the Kullback-Leibler symmetric divergence. A reclassification model is proposed to stabilize the classifier, and the Sammon map is used to visualize the class separation.

Unsupervised color image segmentation based on local fractal descriptor and J-images

This paper proposes an improved version for the JSEG color image segmentation algorithm, combining the classical JSEG algorithm and a local fractal operator that measures the fractal dimension of each pixel, thus improving the boundary detection in the J-map. Experiments with natural color images of the Berkeley Segmentation Dataset and Benchmark are presented, which show improved results in comparison with the classical JSEG algorithm.

Fractal-JSEG: JSEG Using an Homogeneity Measurement Based on Local Fractal Descriptor

This paper proposes an improved version for the JSEG color image segmentation algorithm, combining the classical JSEG algorithm and a local fractal operator that measures the fractal dimension of each pixel, thus improving the boundary detection in the J-map. Experiments with natural color images of the Berkeley Segmentation Dataset and Benchmark are presented, which show improved results in comparison with the classical JSEG algorithm.

Experimental evidences for visual evoked potentials with stimuli beyond the conscious perception threshold

The Steady State Visual Evoked Potentials (SSVEP), present in ElectroEncephaloGram (EEG) signal, are currently used as a convenient approach to a Brain-Computer Interface (BCI). However, the stimulus frequencies are bellow the Flicker Fusion Frequency (FFF). In this work, the possibility of producing SSVEP for stimulus frequency beyond the FFF is investigated. From our experimental SSVEP results, it is shown that there are consistent evidences to support the hypothesis of non-conscious perception. Finally, their practical implications to both engineering and psychological issues are duly discussed.

A pseudo-online Brain-Computer Interface with automatic choice for EEG channel and frequency

This paper presents the classification of three mental tasks, using the electroencephalographic signal and simulating a real-time process, that is, the pseudo-online technique. Linear Discriminant Analysis is used to recognize the mental tasks, and the feature extraction uses the Power Spectral Density. The choice of EEG channel and frequency uses the Kullback-Leibler symmetric divergence and a reclassification model is proposed to stabilize the classifier. Finally, it is expected that the proposed method can be implemented in a Brain-Computer Interface associated with a robotic wheelchair.

Face Tracking in Unconstrained Color Videos with the Recovery of the Location of Lost Faces

This paper proposes a framework to track a face in multi-view uncontrolled color video sequences. The method combines Gabor responses with missing observation Kalman filter to track the face. Using this approach, it is not necessary to estimate face positions even if the detection stage fails, because the missing observation Kalman filter is able to predict the face location in the next frame of the video sequence. Literature shows that tracking approach needs a face observation, returning to initial step. This work uses a preprocessing stage that actively treats the color constancy problem. This algorithm is applied directly to non-normalized RGB space, not demanding any color space transformation. Another contribution is the identification of a range for dark skin tones, not yet identified in uncontrolled color videos. Skin-tone pixel identification reduces the number of candidates to be a face region in the image and ensures that the image region is a human face. Using skin searching region the method can predict the object motion more accurately than one that performs face searching in the whole image. The proposed framework presented encouraging results for both indoor and outdoor unconstrained test videos, considering multi-view scenes containing partial occlusion and non-uniform illumination. Moreover, its capability to recover the face location not detected in a previous frame decreases the whole runtime, making it a very attractive one.

Face detection based on Adaptive Support Vector Tracker

This paper proposes a framework to track faces in color video sequences. The Adaptive Support Vector Tracker (ASVT) combines face detection with target tracking through using an adaptive filter in unconstrained videos. The adjacent locations of the target point are predicted in a search window reducing the number of image regions that are candidates for faces. Thus, the method can predict the object motion more accurately. The architecture is distinguished by the good results for both indoor and outdoor unconstrained videos, for scenes containing scale variation, partial occlusion, bad illumination and complex background. Brightness compensation is applied to improve the detection of faces in videos.