580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can hierarchical multi-resolution control architectures enhance cooperation in multi-agent and multi-robot systems?
This research theme investigates the design and application of hierarchical control architectures that manage cooperation among autonomous agents and robots at multiple levels of abstraction and resolution. The objective is to address the complexity and dynamism of cooperative tasks by decomposing control into layered decision-making processes, balancing global team goals and individual agent autonomy. This approach is key to improving scalability, coordination, and adaptability in cooperative autonomous systems.
Key finding: The MRCC model introduces a three-layer hierarchical control framework—system level, micro-social level, and agent-level individual control—to manage cooperation in multi-robot tasks. The top layers handle team goals and... Read more
Key finding: This work presents a hybrid control architecture featuring a reflective planning agent that translates high-level operator goals into executable missions for co-operating autonomous robots. The system combines abstract... Read more
Key finding: Through cooperative coevolutionary algorithms (CCEAs), this study evolves heterogeneous and specialized control policies for a team of three aquatic surface robots performing cooperative predator-prey pursuit. The approach... Read more
2. How can multi-agent frameworks facilitate cooperative decision-making and active collaboration between autonomous systems and human operators?
This theme focuses on multi-agent system (MAS) architectures and models that enable dynamic collaboration, negotiation, and coordination between autonomous agents and human users. It encompasses research that combines cognitive and control layers to allow intelligent vehicles or software agents to interact actively with humans, adapt their behaviors, modulate alerts or control, and jointly accomplish complex tasks. Understanding how to model teamwork, communication, and shared control in MAS is critical for enhancing autonomous systems that operate within human-centered environments.
Key finding: This work develops a MAS-based interface (OMAS) acting as a digital intermediary between driver and ADAS functions, enabling real-time dialog-based collaboration. Using speech recognition and cognitive-level cooperation, the... Read more
Key finding: This research proposes a hierarchical MAS architecture layered into guidance, management, and traffic control levels for vehicle platoon coordination. It contrasts centralized leader-follower schemes with decentralized... Read more
Key finding: The paper develops a three-layer hierarchical control structure for decentralized vehicle platoon coordination, leveraging multi-agent teamwork models and communication. The decentralized approach treats the platoon as a... Read more
3. What foundational definitions, architectures, and evolution paradigms support cooperation among autonomous agents in multi-agent systems?
This research area investigates the theoretical underpinnings, definitions, typologies, control architectures, and mechanisms of cooperation in multi-agent systems. It includes the characterization of agent autonomy, communication, coordination, negotiation, and the evolution or emergence of cooperative behaviors. Establishing clear conceptual distinctions and architectural frameworks is essential to enable systematic design and implementation of cooperative autonomous systems with predictable and effective interactions.
Key finding: This survey delineates the evolution of multi-agent systems from distributed AI origins, defining software agents as autonomously acting software entities operating in complex environments. It highlights that agents are... Read more
Key finding: This work provides a nuanced definition and typology of cooperation in MAS, distinguishing independent, discrete, and cooperative systems. It articulates forms of cooperation involving communication, deliberation,... Read more
Key finding: The thesis develops a behavior-based distributed control architecture enabling homogeneous agents to plan both individual and cooperative actions uniformly. It emphasizes interaction dynamics, both intra-agent and... Read more
Key finding: This workshop report synthesizes diverse research strands linking mechanisms from economics (mechanism design, auctions) with distributed multi-agent systems, addressing challenges such as modeling rational agent... Read more
Key finding: The paper introduces the DIAMOND method for design and development of embedded hardware/software multi-agent systems, emphasizing the spiral lifecycle combining multi-agent analysis and component-based design. It argues for... Read more