Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

McMaster University

$31,000.00

Certifiable Algorithms for Safe Robot Perception

EF 2024-25

Résumé
Robotic systems are rapidly expanding their historical role as largely single-purpose manufacturing tools into novel application domains that require a far greater degree of complexity. Examples of this trend include the rapid development of self-driving cars, the proliferation of micro aerial vehicles, and the increasing presence of general-purpose collaborative robots in warehouses and other settings shared with humans. In order to fully realize the immense economic and scientific potential of these systems, roboticists must earn the confidence of many diverse stakeholders: policymakers, investors, and most importantly, the general public which autonomous robots will serve. To this end, robotics researchers need to equip the next generation of mobile robots with some of the safety guarantees and structure characteristic of traditional disciplines such as civil engineering and aviation. One common pitfall for autonomous mobile robots is inaccurate or completely erroneous perception of their environment (i.e., their map) and their pose relative to this map. Without trustworthy mapping and localization, safe and correct completion of tasks is impossible. While many diverse solutions to this problem of spatial awareness exist, dealing with complex dynamic environments, multiple agents, and extreme sensor noise or faults remains challenging. In fact, without very accurate sensor measurements, intervention by a human operator, or an accurate prior estimate, state-of-the-art localization methods are liable to return highly suboptimal solutions. Some of these methods even incorporate deep perception models trained on large datasets, complicating the detection and interpretability of estimation errors and other faults. To address this shortcoming, the proposed research aims to develop techniques for finding certifiably globally optimal solutions to perception problems faced by autonomous mobile robots. In the next five years, this will involve broadening the current theoretical understanding of global optimization for perception, applying these methods to complex multi-sensor calibration problems, and developing and testing a software toolchain for automatically generating efficient code from high-level descriptions of perception problems. This undertaking will lead to accessible open-source software that allows the broader robotics community to use certifiable methods on a variety of robot platforms suitable for numerous applications. If deployed equitably, autonomous mobile robots have the potential to transform Canada's workforce and economy by easing the burden of dangerous and dull tasks, improving productivity and the health and safety of workers. Finally, the global climate change crisis has made preserving and understanding our nation's vast natural resources an urgent priority. Certifiable perception for ground, aerial, and marine robots will enable environmental monitoring at a scale which is transformative for climate science.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2023-04836
Name nom
Giamou, Matthew
Department département
Computing and Software
Organization id
27
Institution établissement
McMaster University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
31000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
1206
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Engineering
Area of application fr
Génie
Research subject code
2600
Research subject group en
Robotics
Research subject group fr
Robotique
Research subject en
Robotics
Research subject fr
Robotique
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Certifiable Algorithms for Safe Robot Perception
Keywords
certifiability; optimization; safety; state estimation; convexity; sensor calibration; autonomous navigation; deep learning; perception; robot vision
Application summary
Robotic systems are rapidly expanding their historical role as largely single-purpose manufacturing tools into novel application domains that require a far greater degree of complexity. Examples of this trend include the rapid development of self-driving cars, the proliferation of micro aerial vehicles, and the increasing presence of general-purpose collaborative robots in warehouses and other settings shared with humans. In order to fully realize the immense economic and scientific potential of these systems, roboticists must earn the confidence of many diverse stakeholders: policymakers, investors, and most importantly, the general public which autonomous robots will serve. To this end, robotics researchers need to equip the next generation of mobile robots with some of the safety guarantees and structure characteristic of traditional disciplines such as civil engineering and aviation. One common pitfall for autonomous mobile robots is inaccurate or completely erroneous perception of their environment (i.e., their map) and their pose relative to this map. Without trustworthy mapping and localization, safe and correct completion of tasks is impossible. While many diverse solutions to this problem of spatial awareness exist, dealing with complex dynamic environments, multiple agents, and extreme sensor noise or faults remains challenging. In fact, without very accurate sensor measurements, intervention by a human operator, or an accurate prior estimate, state-of-the-art localization methods are liable to return highly suboptimal solutions. Some of these methods even incorporate deep perception models trained on large datasets, complicating the detection and interpretability of estimation errors and other faults. To address this shortcoming, the proposed research aims to develop techniques for finding certifiably globally optimal solutions to perception problems faced by autonomous mobile robots. In the next five years, this will involve broadening the current theoretical understanding of global optimization for perception, applying these methods to complex multi-sensor calibration problems, and developing and testing a software toolchain for automatically generating efficient code from high-level descriptions of perception problems. This undertaking will lead to accessible open-source software that allows the broader robotics community to use certifiable methods on a variety of robot platforms suitable for numerous applications. If deployed equitably, autonomous mobile robots have the potential to transform Canada's workforce and economy by easing the burden of dangerous and dull tasks, improving productivity and the health and safety of workers. Finally, the global climate change crisis has made preserving and understanding our nation's vast natural resources an urgent priority. Certifiable perception for ground, aerial, and marine robots will enable environmental monitoring at a scale which is transformative for climate science.