Exploration and characterization of environments under uncertainty using multimodal sensors

• Autores: Carlos García Saura
• Directores de la Tesis: Pablo Varona (dir. tes.) , Francisco de Borja Rodríguez Ortiz (dir. tes.)
• Lectura: En la Universidad Autónoma de Madrid ( España ) en 2023
• Idioma: inglés
• Número de páginas: 137
• Títulos paralelos:
  • Exploración y caracterización de entornos con incertidumbre mediante sensores multimodales
• Tribunal Calificador de la Tesis: Jesus Tornero Lopez (presid.) , Manuel Sánchez-Montañés Isla (secret.) , Juan González Gómez (voc.)
• Programa de doctorado: Programa de Doctorado en Ingeniería Informática y de Telecomunicación por la Universidad Autónoma de Madrid
• Calificación: Sobresaliente Cum Laude
• Enlaces
  • Tesis en acceso abierto en: Biblos-e Archivo
• Resumen

  • This thesis explores the field of artificial noses and autonomous robotic search strategies in environments with uncertainty. The focus has been set on the development of artificial noses both in fixed locations and on mobile robots. In fixed locations, the use of artificial noses has shown the potential to improve the accuracy and efficiency of various processes, such as food and beverage quality control, noninvasive human activity monitoring, and air quality assessment. On mobile robots, the use of these artificial noses together with robust search strategies effectively enhances their capabilities, making them useful in various applications, including search and rescue missions, agriculture, and environmental monitoring. We have developed a wireless olfactory sensor platform and demonstrated its long-term robustness in various real-world environments. Our platform has acquired real-time multiparameter data from a set of classrooms at our university since 2016, with the goal of monitoring and characterizing human routines and, for the first time, cognitive human activity. The obtained time series datasets are of interest as the recording started a long time before the COVID-19 pandemic and subsequent changes in ventilation regulation and educational protocols. In parallel, we have also performed multiple experiments related to the food industry, to evaluate the capacity of these sensors to characterize and certify the evolution of alimentary processes. Additionally, we analyzed the factors that influence autonomous robotic search in uncertain environments such as odor localization and demonstrated how these factors can be utilized to improve the performance of exploratory search algorithms. For a comprehensive analysis we implemented a robot simulation platform calibrated to our real-world experiments. The proposed characterization approach allows in turn to design robots better adapted to the particularities of each environment by considering the knowledge, even if restricted, of the search area, specific motor plant characteristics such as motion drift, sensor range, obstacle handling, battery life, etc. We also explored a possible approach for incorporating these algorithms into an underwater robot, by designing a bio-inspired electrolocation sensor. This could enable accurate obstacle detection and in turn the implementation of our search strategies, to provide more autonomy in real-world scenarios. Our work has resulted in journal and conference publications, and two patents that describe novel applications for olfactory sensors, such as non-invasive monitoring of human cognitive activity by registering different parameters in ambient air, and to better characterize the food supply chain by evaluating and certifying odor profile evolutions over the product lifespan.