Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

Carleton University

$36,000.00

Toward the Use of the Belief - Desire - Intention Paradigm in Autonomous Mobile Robots and Vehicles

EF 2024-25

Résumé
While the recent gains in machine learning techniques and especially deep neural networks has allowed major progress in many areas of robotics (convolutional neural networks (CNN) in vision, deep reinforcement learning (RL) in control), there is still a need to program the mission of mobile robots in a way that is easy for developers and makes the resulting behaviour understandable yet highly efficient. The Belief-Desire-Intention (BDI) paradigm is specifically meant to address issues of autonomous behaviour in agents. Indeed, BDI allows the agent developer to specify the agent or robot's mission explicitly in terms of goals, plans and contingencies, and a BDI framework such as AgentSpeak will select, execute and monitor the plans in the appropriate context. However, for BDI to be successful in such applications, further work is required. Our research program is in addressing the suitability of BDI in autonmous mobile robots and vehicles, and consists of improving the ease of development, expressiveness, performance and explainability issues. More specifically: - currently, an AgentSpeak developer needs to specify BDI plans "by hand" in hard-coded fashion. We propose to investigate the use of Linear Planning, so that the developer only needs to specify the pre- and post-conditions of the operators available to the agent, and let the agent itself string together the operators in order to reach its goals. - BDI has mostly been investigated in simulations where there isn't much requirement to handle uncertainty. We propose to increase the expressiveness of the AgentSpeak language semantics to better specify plans where the knowledge of the surrounding environment is partial or unreliable. - the above enhancements will have a cost in terms of performance. We propose enhancing existing BDI implementations with a unified high-performance search-based solver that can be used in all performance bottlenecks such as the linear planner, the evaluation of plan guards and the reasoner for handling uncertainty. - an autonomous robotic system would not be complete without a way to easily test, debug and ultimately understand the behaviour of the agent. BDI by definition exposes the mental model of the agent and helps answer "why" questions. We propose investigating the use of a BDI "black box" that will allow us to interrogate and verify that the agent behaviour was consistent with its mission and fullfilled safety and ethical properties. - we will validate our approach by integrating our enhancements with a full robotics control system based on ROS and lower-layer perception, action and control subsystems. The resulting framework will be evaluated in various case studies involving mobile robotic missions such as autonomous mail-delivery, bridge inspection, and autonomous driving.
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-2023-05102
Name nom
Esfandiari, Babak
Department département
Systems and Computer Engineering
Organization id
23
Institution établissement
Carleton 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
36000
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
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2705
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Software and development
Research subject fr
Logiciels et développement
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
Toward the Use of the Belief - Desire - Intention Paradigm in Autonomous Mobile Robots and Vehicles
Keywords
agent-based systems; belief-desires-intentions; autonomous vehicles
Application summary
While the recent gains in machine learning techniques and especially deep neural networks has allowed major progress in many areas of robotics (convolutional neural networks (CNN) in vision, deep reinforcement learning (RL) in control), there is still a need to program the mission of mobile robots in a way that is easy for developers and makes the resulting behaviour understandable yet highly efficient. The Belief-Desire-Intention (BDI) paradigm is specifically meant to address issues of autonomous behaviour in agents. Indeed, BDI allows the agent developer to specify the agent or robot's mission explicitly in terms of goals, plans and contingencies, and a BDI framework such as AgentSpeak will select, execute and monitor the plans in the appropriate context. However, for BDI to be successful in such applications, further work is required. Our research program is in addressing the suitability of BDI in autonmous mobile robots and vehicles, and consists of improving the ease of development, expressiveness, performance and explainability issues. More specifically: - currently, an AgentSpeak developer needs to specify BDI plans "by hand" in hard-coded fashion. We propose to investigate the use of Linear Planning, so that the developer only needs to specify the pre- and post-conditions of the operators available to the agent, and let the agent itself string together the operators in order to reach its goals. - BDI has mostly been investigated in simulations where there isn't much requirement to handle uncertainty. We propose to increase the expressiveness of the AgentSpeak language semantics to better specify plans where the knowledge of the surrounding environment is partial or unreliable. - the above enhancements will have a cost in terms of performance. We propose enhancing existing BDI implementations with a unified high-performance search-based solver that can be used in all performance bottlenecks such as the linear planner, the evaluation of plan guards and the reasoner for handling uncertainty. - an autonomous robotic system would not be complete without a way to easily test, debug and ultimately understand the behaviour of the agent. BDI by definition exposes the mental model of the agent and helps answer "why" questions. We propose investigating the use of a BDI "black box" that will allow us to interrogate and verify that the agent behaviour was consistent with its mission and fullfilled safety and ethical properties. - we will validate our approach by integrating our enhancements with a full robotics control system based on ROS and lower-layer perception, action and control subsystems. The resulting framework will be evaluated in various case studies involving mobile robotic missions such as autonomous mail-delivery, bridge inspection, and autonomous driving.