3 research outputs found
Empirical comparison of methods for the design of robot swarms: a simulation study of swarms that coordinate other swarms.
No full textilustraciones, tablas.Esta tesis presenta una comparación empírica de métodos de diseño de comportamientos colectivos para enjambres de robots que deben coordinar a otros enjambres. En este estudio, los métodos de diseño producen software de control sin ser provistos con información explícita sobre el comportamiento de los enjambres que deben ser coordinados—esta información debe ser obtenida durante el proceso de diseño. Los métodos considerados en la comparación son Pistacchio, EvoCMY, C-Human, R-Walk. Pistacchio es un método automático modular, perteneciente a la familia de métodos AutoMoDe. EvoCMY es un método automático basado en neuroevolución. La eficiencia de los métodos automáticos fue comparada con C-Human, un método de diseño manual en donde un grupo de diseñadores producen el software de control para los robots. Pistacchio, EvoCMY y C-Human fueron comparados con R-Walk, una implementación de movimiento aleatorio que sirve como línea base. Pistacchio y EvoCMY se presentan por primera vez en esta tesis, y son una contribución de la investigación. Los experimentos llevados a cabo con Pistacchio y EvoCMY permitieron determinar que los métodos de diseño automático son efectivos en (i) la identificación de posibles formas de interacción entre el enjambre que coordina y el enjambre que es coordinado; y en (ii) el aprovechamiento de estas formas de interacción para ejecutar misiones en donde los dos enjambres operan de forma coordinada. La comparación de los métodos se llevó a cabo empleando ARGoS3, un simulador especializado en robótica de enjambres, y tomando como plataforma de referencia al robot e-puck. El diseño mediante métodos automáticos resultó significativamente más eficiente que el diseño manual en las misiones propuestas (Texto tomado de la fuente).This thesis presents an empirical comparison of methods for the design of collective behaviors for a robot swarm that must coordinate a second robot swarm. In this study, the design methods produce control software without being provided with explicit information about the behavior of the swarm that must be coordinated—in all cases, this information must be obtained during the design process. The methods considered in the study are Pistacchio, EvoCMY, C-Human, R-Walk. Pistacchio is an automatic modular method that belongs to the family of methods AutoMoDe. EvoCMY is an automatic method based on neuro-evolution. The performance of the automatic methods was compared with the performance of C-Human, a manual design method in which a group of human designers produces the control software for the robots. Pistacchio, EvoCMY and C-Human were compared against R-Walk, an implementation of a random walk that serves as a baseline. Pistacchio and EvoCMY are presented for the first time in this thesis, and they are an original contribution. Experiments conducted with Pistacchio and EvoCMY showed that automatic design methods are effective in (i) identifying interaction dynamics between the swarm that coordinates and the swarm that is being coordinated; and in (ii) using these dynamics for addressing missions in which the two robot swarms must act in a coordinated manner. The design methods were compared in simulation, using the swarm-dedicated simulator ARGoS3 and a simulated model of the e-puck robot. The automatic design approach performed significantly better than the manual design in the proposed missions.MaestríaMagíster en Ingeniería - Automatización IndustrialDiseño de comportamientos colectivos para enjambres de robotsEléctrica, Electrónica, Automatización Y Telecomunicacione
Implementation of a Brazilian Cardioprotective Nutritional (BALANCE) Program for improvement on quality of diet and secondary prevention of cardiovascular events: A randomized, multicenter trial
No full textBackground: Appropriate dietary recommendations represent a key part of secondary prevention in cardiovascular disease (CVD). We evaluated the effectiveness of the implementation of a nutritional program on quality of diet, cardiovascular events, and death in patients with established CVD. Methods: In this open-label, multicenter trial conducted in 35 sites in Brazil, we randomly assigned (1:1) patients aged 45 years or older to receive either the BALANCE Program (experimental group) or conventional nutrition advice (control group). The BALANCE Program included a unique nutritional education strategy to implement recommendations from guidelines, adapted to the use of affordable and regional foods. Adherence to diet was evaluated by the modified Alternative Healthy Eating Index. The primary end point was a composite of all-cause mortality, cardiovascular death, cardiac arrest, myocardial infarction, stroke, myocardial revascularization, amputation, or hospitalization for unstable angina. Secondary end points included biochemical and anthropometric data, and blood pressure levels. Results: From March 5, 2013, to Abril 7, 2015, a total of 2534 eligible patients were randomly assigned to either the BALANCE Program group (n = 1,266) or the control group (n = 1,268) and were followed up for a median of 3.5 years. In total, 235 (9.3%) participants had been lost to follow-up. After 3 years of follow-up, mean modified Alternative Healthy Eating Index (scale 0-70) was only slightly higher in the BALANCE group versus the control group (26.2 ± 8.4 vs 24.7 ± 8.6, P <.01), mainly due to a 0.5-serving/d greater intake of fruits and of vegetables in the BALANCE group. Primary end point events occurred in 236 participants (18.8%) in the BALANCE group and in 207 participants (16.4%) in the control group (hazard ratio, 1.15; 95% CI 0.95-1.38; P =.15). Secondary end points did not differ between groups after follow-up. Conclusions: The BALANCE Program only slightly improved adherence to a healthy diet in patients with established CVD and had no significant effect on the incidence of cardiovascular events or death. © 2019 The Author
Gendered self-views across 62 countries: A test of competing models
No full textSocial role theory posits that binary gender gaps in agency and communion should be larger in less egalitarian countries, reflecting these countries’ more pronounced sex-based power divisions. Conversely, evolutionary and self-construal theorists suggest that gender gaps in agency and communion should be larger in more egalitarian countries, reflecting the greater autonomy support and flexible self-construction processes present in these countries. Using data from 62 countries (N = 28,640), we examine binary gender gaps in agentic and communal self-views as a function of country-level objective gender equality (the Global Gender Gap Index) and subjective distributions of social power (the Power Distance Index). Findings show that in more egalitarian countries, gender gaps in agency are smaller and gender gaps in communality are larger. These patterns are driven primarily by cross-country differences in men’s self-views and by the Power Distance Index (PDI) more robustly than the Global Gender Gap Index (GGGI). We consider possible causes and implications of these findings.<br/
