Enhancement of the Sensory Capabilities of Mobile Robots through Artificial Olfaction
| Type of publication: | Phdthesis |
| Citation: | |
| Year: | 2021 |
| Month: | {{jan}} |
| School: | Universidad de Málaga |
| URL: | http://mapir.uma.es/papersrepo... |
| Abstract: | This thesis covers several aspects of artificial or robotic olfaction, that is, sensing information about the composition of the air surrounding an automated system. First, we have developed an electronic nose, an instrument that combines low-performance gas sensors with a classification algorithm to measure and identify gases by their chemical signature. Although this technology already existed previously, a new approach is applied that seeks to reduce dimensions and consumption in order to be installed in mobile robots, whilst also increasing the number of detectable gases by means of a modular design. Subsequently, we have explored a possible optimal strategy for finding gas leaks with a robot equipped with this type of electronic noses. In particular, we have carried out several experiments based on teleoperation to understand how the robot sensors affect the success of the task, allowing us to derive an algorithm to generate gas-maps (i.e. physical gas distribution maps) with autonomous robots based on the observed human search behavior: maximizing the known information about the environment. The main virtue of this method, besides performing an optimal exploration of the environment, is its ability to operate in very complex environments and subject to wind currents using a new method that is also presented in this thesis. Finally, two application cases are also presented in which we automatically identify and qualify the subjective air quality in urban environments. La presente tesis abarca varios aspectos del olfato artificial u olfato robótico, la capacidad de percibir información sobre la composición del aire que rodea a un sistema automático. En primer lugar, se desarrolla una nariz electrónica, un instrumento que combina sensores de gas de bajas prestaciones con un algoritmo de clasificación para medir e identificar gases. Aunque esta tecnología ya existía previamente, se aplica un nuevo enfoque que busca reducir las dimensiones y consumo para poder instalarlas en robots móviles, a la vez que se aumenta el número de gases detectables mediante un diseño modular. Posteriormente, se estudia la estrategia óptima para encontrar fugas de gas con un robot equipado con este tipo de narices electrónicas. Para ello se llevan a cabos varios experimentos basados en teleoperación para entender como afectan los sensores del robot al éxito de la tarea, de lo cual se deriva finalmente un algoritmo para generar con robots autónomos mapas de gas de un entorno dado, el cual se inspira en el comportamiento humano, a saber, maximizar la información conocida sobre el entorno. La principal virtud de este método, además de realizar una exploración óptima del entorno, es su capacidad para funcionar en entornos muy complejos y sujetos a corrientes de vientos mediante un nuevo método que también se presenta en esta tesis. Finalmente, se presentan dos casos de aplicación en los que se identifica de forma automática con una nariz electrónica la calidad subjetiva del aire en entornos urbanos. |
| Userfields: | img_url=http%3A%2F%2Fmapir.isa.uma.es%2Fimagesrepo%2Fthesis%2F2021-andresgongora.png,rank_indexname=,rank_pos_in_category=,rank_num_in_category=,rank_cat_name=,impact_factor= |
| Keywords: | |
| Authors | |
| Added by: | [] |
| Total mark: | 0 |
|
Attachments
|
|
|
|
|
Notes
|
|
|
|
|
|
Topics
|
|
|
|
|
