Grant: Research Tools and Instruments
Université de Montréal
$150,000.00
Generalization and Planning in Robotics
EF 2023-24
Résumé
The Robotics and Embodied AI Lab (REAL) is in the Département d'Informatique et de Recherche Opérationnelle (DIRO) at the Université de Montréal and is also affiliated with Mila. REAL is directed by professors Liam Paull and Glen Berseth and has 4 research assistants, over 20 students and is affiliated with 4 professors around Montreal. The research in the lab uses machine learning methods to enable robots to complete complex real-world tasks, such as cleaning, cooking, and manufacturing among others. Research in this area has reached a level of maturity that requires diverse and complex environments for proper and rigorous evaluation. The PI (Berseth) and Co-PI (Paull) are experts in learning-based methods for robotics and are extremely well-positioned to study, evaluate and develop world-class robotics methods. However, their current lab lacks (1) the diverse environments and the newly created (2) mobile robots that can act in these very human-centric-designed environments. The lack of mobile manipulators that have human height greatly restricts the type of problems, algorithms and diversity we can explore. Until recently, artificially constructed university "toy" environments could be used as proxies for complex real-world environments in which robots would actually be deployed. However, we recently see the emergence of robot demonstrations in actual large-scale realistic environments. Our lack of hardware and infrastructure is truly limiting our labs' ability to stay at the forefront of the research space, which has been quickly advancing into problems that involve mobile manipulation [SayCan, SayPlan, etc]. There will be many years of research needed in this area. In order to compete in this space, it becomes a requirement to construct a much more realistic environment for data-collection, interaction, and evaluation. Why a kitchen?: A kitchen is an appropriate environment since it contains a high diversity of objects which we interact with in a diversity of ways. Operating across the kitchens shown in Figure 1, requires problem-solving and reasoning at multiple levels, from fine-grained manipulation to high-level reasoning about cooking times and environmental organization. In addition, some students have started trying to collect data from kitchen-like environments with stationary robot arms, but the kitchens are often in use, dangerous to use with real hardware, because of liquids, and not realistic or diverse. We aim to build two almost real kitchens from Ikea that will not require construction to install but provide diverse motion planning problems. Between Liam and Glen, there is significant funding for students but we are limited in providing the students with sufficient hardware to enable the best possible HQP. Robotics is a priority area for Canada and CIFAR and the lab aims to train students to take roles to pursue complex problems in the space of more flexible automation methods, cooking robots, and surveying.
Programme
Research Tools and Instruments
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RTI-2024-00182
Name nom
Berseth, Glen
Department département
DIRO (Computer Science)
Organization id
63
Institution établissement
Université de Montréal
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2024
Award amount
150000
Program id
RTI
Program name en
Research Tools and Instruments
Program name fr
Outils et instruments de recherche
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1607
Committee name en
Computer, Mathematical and Statistical Sciences (RTI)
Committee name fr
Sciences informatiques, mathématiques et statistiques (OIR)
Area of application code
500
Area of application group en
Manufacturing processes and products
Area of application group fr
Procédés et produits industriels
Area of application en
Manufacturing processes and products
Area of application fr
Procédés et produits industriels
Research subject code
2805
Research subject group en
Artificial intelligence
Research subject group fr
Learning and inference theories
Research subject en
Intelligence artificielle
Research subject fr
Théories de l'apprentissage et l'inférence
Application title
Generalization and Planning in Robotics
Keyword
machine learning; manipulation; perception; planning; robotics
Application summary
The Robotics and Embodied AI Lab (REAL) is in the Département d'Informatique et de Recherche Opérationnelle (DIRO) at the Université de Montréal and is also affiliated with Mila. REAL is directed by professors Liam Paull and Glen Berseth and has 4 research assistants, over 20 students and is affiliated with 4 professors around Montreal. The research in the lab uses machine learning methods to enable robots to complete complex real-world tasks, such as cleaning, cooking, and manufacturing among others. Research in this area has reached a level of maturity that requires diverse and complex environments for proper and rigorous evaluation. The PI (Berseth) and Co-PI (Paull) are experts in learning-based methods for robotics and are extremely well-positioned to study, evaluate and develop world-class robotics methods. However, their current lab lacks (1) the diverse environments and the newly created (2) mobile robots that can act in these very human-centric-designed environments. The lack of mobile manipulators that have human height greatly restricts the type of problems, algorithms and diversity we can explore. Until recently, artificially constructed university "toy" environments could be used as proxies for complex real-world environments in which robots would actually be deployed. However, we recently see the emergence of robot demonstrations in actual large-scale realistic environments. Our lack of hardware and infrastructure is truly limiting our labs' ability to stay at the forefront of the research space, which has been quickly advancing into problems that involve mobile manipulation [SayCan, SayPlan, etc]. There will be many years of research needed in this area. In order to compete in this space, it becomes a requirement to construct a much more realistic environment for data-collection, interaction, and evaluation. Why a kitchen?: A kitchen is an appropriate environment since it contains a high diversity of objects which we interact with in a diversity of ways. Operating across the kitchens shown in Figure 1, requires problem-solving and reasoning at multiple levels, from fine-grained manipulation to high-level reasoning about cooking times and environmental organization. In addition, some students have started trying to collect data from kitchen-like environments with stationary robot arms, but the kitchens are often in use, dangerous to use with real hardware, because of liquids, and not realistic or diverse. We aim to build two almost real kitchens from Ikea that will not require construction to install but provide diverse motion planning problems. Between Liam and Glen, there is significant funding for students but we are limited in providing the students with sufficient hardware to enable the best possible HQP. Robotics is a priority area for Canada and CIFAR and the lab aims to train students to take roles to pursue complex problems in the space of more flexible automation methods, cooking robots, and surveying.