Skip to main content

Maria Gini

University of Minnesota, Computer Science and Engineerng, Faculty Member

Followers

139

Following

34

Co-authors

14

Mentions

1

Public Views

I am a Professor of Computer Science at the University of Minnesota. My research is on distributed decision making for autonomous agents. I work in different application areas, ranging from distributed methods for allocation of tasks, to methods to explore an unknown environment, teamwork for search and rescue, and navigation in dense crowds.
Address: Computer Science and Engineering, University of Minnesota, Minneapolis, Minnesota, United States

less

InterestsView All (12)

Uploads

Papers by Maria Gini

Connectionist reinforcement learning for control of robotic systems

Flexible Decision Support in Dynamic Interorganizational Networks

Lazy max-sum for allocation of tasks with growing costs

Robotics and Autonomous Systems, Dec 1, 2018

We propose a model for the allocation of agents to tasks when the tasks have a cost which grows o... more We propose a model for the allocation of agents to tasks when the tasks have a cost which grows over time. Our model accounts for both the natural growth of tasks and the effort of the agents at containing such growth. The objective is to produce solutions that minimize the growth of tasks (potentially stopping such growth) by efficiently coordinating the operations of the agents. This problem has strong spatial and temporal components, as the agents require time not only to work on the tasks but also to move between tasks and during that time the costs of completing the tasks continue to grow. We propose a novel distributed coordination algorithm, called Lazy max-sum, which works well even when the model of the environment has errors. The algorithm handles homogeneous as well as heterogeneous agents, which can do different amounts of work per time unit and have different travel speeds. We show experimentally that the algorithm outperforms other methods in both a simple simulation and the RoboCup Rescue agent simulation.

<title>Performance of a distributed robotic system using shared communications channels</title>

Proceedings of SPIE, Oct 4, 2001

Flexible and Interactive Decision Support in a Dynamic Business Network

Programming Vision Systems and Integrating them with Robots

IFAC Proceedings Volumes, Nov 1, 1985

I-Ie p r esent LIVIA, a language for writing application programs for industrial vision systems. ... more I-Ie p r esent LIVIA, a language for writing application programs for industrial vision systems. LIVIA is a Pascal-like language wi t h specific ins tr uctions fo r vision. The objects are identified by the values of their invariant features. Operations on the objects can be done using the elementary set theory. The r ecognit i on of objects in a data base and the selection of any of their features is easily programmed.

Error Diagnosis and Repair in Sensorial Robots

Control of Intelligent Robots and Goal Orientated Languages

Industrial Robot-an International Journal, 1975

Using a Task Description Language for Assembly. The Generation of World Models

Constructing Programs with a Plan Formation System

IFAC Proceedings Volumes, Oct 1, 1983

This paper describes a robot planning system for constructing manipulator programs from task desc... more This paper describes a robot planning system for constructing manipulator programs from task descriptions. Plan formation and program generation techniques are used to transform task specifications into executable programs. The plan formation operates on a data bas e of knowledge a bo~ the world to produce the sequence of actions needed to execute the task. T~, program generation transforms the sequence of actions into an executable pro!V'am, written in MAL, using knowledge about the robot and its control syS'tem.

Detecting and forecasting economic regimes in multi-agent automated exchanges

Decision Support Systems, Nov 1, 2009

We show how an autonomous agent can use observable market conditions to characterize the microeco... more We show how an autonomous agent can use observable market conditions to characterize the microeconomic situation of the market and predict future market trends. The agent can use this information to make both tactical decisions, such as pricing, and strategic decisions, such as product mix and production planning. We develop methods to learn dominant market conditions, such as oversupply or scarcity, from historical data using Gaussian mixture models to construct price density functions. We discuss how this model can be combined with real-time observable information to identify the current dominant market condition and to forecast market changes over a planning horizon. We forecast market changes via both a Markov correction-prediction process and an exponential smoother. Empirical analysis shows that the exponential smoother yields more accurate predictions for the current and the next day (supporting tactical decisions), while the Markov correction-prediction process is better for longer term predictions (supporting strategic decisions). Our approach offers more flexibility than traditional regression based approaches, since it does not assume a fixed functional relationship between dependent and independent variables. We validate our methods by presenting experimental results in a case study, the Trading Agent Competition for Supply Chain Management.

Decision Processes in Agent-Based Contracting

The Magnet system meets many of the challenges of modeling decision making for customer agents in... more

Detecting and Forecasting Economic Regimes in Automated Exchanges

We present basic building blocks of an agent that can use observable market conditions to charact... more We present basic building blocks of an agent that can use observable market conditions to characterize the microeconomic conditions of the market and predict future market trends. The agent can use this information to make both tactical decisions such as pricing and strategic decisions such as product mix and production planning. We develop methods that can learn dominant market conditions, such as oversupply or scarcity, from historical data using computational methods to construct price density functions. We discuss how this knowledge can be used, together with real-time observable information, to identify the current dominant market condition and to forecast market changes over a planning horizon. We validate our methods by presenting experimental results in a case study, the Trading Agent Competition for Supply Chain Management.

Multi-Robot Auctions for Allocation of Tasks with Temporal Constraints

Proceedings of the AAAI Conference on Artificial Intelligence

We propose an auction algorithm to allocate tasks that have temporal constraints to cooperative r... more We propose an auction algorithm to allocate tasks that have temporal constraints to cooperative robots. Temporal constraints are expressed as time windows, within which a task must be executed. There are no restrictions on the time windows, which are allowed to overlap. Robots model their temporal constraints using a simple temporal network, enabling them to maintain consistent schedules. When bidding on a task, a robot takes into account its own current commitments and an optimization objective, which is to minimize the time of completion of the last task alone or in combination with minimizing the distance traveled. The algorithm works both when all the tasks are known upfront and when tasks arrive dynamically. We show the performance of the algorithm in simulation with different numbers of tasks and robots, and compare it with a baseline greedy algorithm and a state-of-the-art auction algorithm. Our algorithm is computationally frugal and consistently allocates more tasks than th...

Implicit Coordination in Crowded Multi-Agent Navigation

Proceedings of the AAAI Conference on Artificial Intelligence

In crowded multi-agent navigation environments, the motion of the agents is significantly constra... more In crowded multi-agent navigation environments, the motion of the agents is significantly constrained by the motion of the nearby agents. This makes planning paths very difficult and leads to inefficient global motion. To address this problem, we propose a new distributed approach to coordinate the motions of agents in crowded environments. With our approach, agents take into account the velocities and goals of their neighbors and optimize their motion accordingly and in real-time. We experimentally validate our coordination approach in a variety of scenarios and show that its performance scales to scenarios with hundreds of agents.

Monte Carlo Tree Search for Multi-Robot Task Allocation

Proceedings of the AAAI Conference on Artificial Intelligence

Multi-robot teams are useful in a variety of task allocation domains such as warehouse automation... more Multi-robot teams are useful in a variety of task allocation domains such as warehouse automation and surveillance. Robots in such domains perform tasks at given locations and specific times, and are allocated tasks to optimize given team objectives. We propose an efficient, satisficing and centralized Monte Carlo TreeSearch based algorithm exploiting branch and bound paradigm to solve the multi-robot task allocation problem with spatial, temporal and other side constraints. Unlike previous heuristics proposed for this problem, our approach offers theoretical guarantees and finds optimal solutions for some non-trivial data sets.

Adaptive Learning for Multi-Agent Navigation

Adaptive Agents and Multi-Agents Systems, May 4, 2015

When agents in a multi-robot system move, they need to adapt their paths to account for potential... more When agents in a multi-robot system move, they need to adapt their paths to account for potential collisions with other agents and with static obstacles. Existing distributed navigation methods compute motions that are optimal locally but do not account for the aggregate motions of all the agents. When there are many agents that move in a crowded space, the result is an inefficient global behavior. To address this issue, we propose a new approach which leverages techniques from machine learning and game theory. Agents using our approach dynamically adapt their motion depending on local conditions in their current environment. We validate our approach experimentally in a variety of scenarios and with different numbers of agents. When compared to other machine learning techniques, our approach produces motions that are more efficient and make better use of the space, allowing agents to reach their destinations faster.

Measuring Physiological Markers of Stress During Conversational Agent Interactions

AI for Disease Surveillance and Pandemic Intelligence, 2022

Moving in a Crowd: Safe and Efficient Navigation among Heterogeneous Agents

Multi-agent navigation methods typically assume that all agents use the same underlying framework... more Multi-agent navigation methods typically assume that all agents use the same underlying framework to navigate to their goal while avoiding colliding with each other. However, such assumption does not hold when agents do not know how other agents will move. We address this issue by proposing a Bayesian inference approach where an agent estimates the navigation model and goal of each neighbor, and uses this to compute a plan that minimizes collisions while driving it to its goal. Simulation experiments performed in many scenarios demonstrate that an agent using our approach computes safer and more time-efficient paths as compared to those generated without our inference approach and a state-of-the-art local navigation framework.

Graphically Representing Child-Robot Interaction Proxemics

This paper discusses the current analysis of a large set of child-robot interaction experiments, ... more This paper discusses the current analysis of a large set of child-robot interaction experiments, in which a 2–4 year old child plays a set of games with a small humanoid robot, with the goal of detecting early signs of autism in toddlers based on the interactions. Our first goal in this paper is to condense and display these child-robot interactions as multi-channel time series, starting with the distances between the child, robot, parent, and experimenter. Our second goal is to use these data displays to compare and contrast different children, with the aim of clustering children with similar interaction patterns. Using a ceiling-mounted camera, we record the interaction between a child and a robot which performs different games and dances. After analyzing the video footage for the locations of all people and the robot in the room over the variable length of the interaction, we condense the interactions into simplified, quantifiable, scale-invariant data. We show how the distances ...

Connectionist reinforcement learning for control of robotic systems

Flexible Decision Support in Dynamic Interorganizational Networks

Lazy max-sum for allocation of tasks with growing costs

Robotics and Autonomous Systems, Dec 1, 2018

We propose a model for the allocation of agents to tasks when the tasks have a cost which grows o... more We propose a model for the allocation of agents to tasks when the tasks have a cost which grows over time. Our model accounts for both the natural growth of tasks and the effort of the agents at containing such growth. The objective is to produce solutions that minimize the growth of tasks (potentially stopping such growth) by efficiently coordinating the operations of the agents. This problem has strong spatial and temporal components, as the agents require time not only to work on the tasks but also to move between tasks and during that time the costs of completing the tasks continue to grow. We propose a novel distributed coordination algorithm, called Lazy max-sum, which works well even when the model of the environment has errors. The algorithm handles homogeneous as well as heterogeneous agents, which can do different amounts of work per time unit and have different travel speeds. We show experimentally that the algorithm outperforms other methods in both a simple simulation and the RoboCup Rescue agent simulation.

<title>Performance of a distributed robotic system using shared communications channels</title>

Proceedings of SPIE, Oct 4, 2001

Flexible and Interactive Decision Support in a Dynamic Business Network

Programming Vision Systems and Integrating them with Robots

IFAC Proceedings Volumes, Nov 1, 1985

I-Ie p r esent LIVIA, a language for writing application programs for industrial vision systems. ... more I-Ie p r esent LIVIA, a language for writing application programs for industrial vision systems. LIVIA is a Pascal-like language wi t h specific ins tr uctions fo r vision. The objects are identified by the values of their invariant features. Operations on the objects can be done using the elementary set theory. The r ecognit i on of objects in a data base and the selection of any of their features is easily programmed.

Error Diagnosis and Repair in Sensorial Robots

Control of Intelligent Robots and Goal Orientated Languages

Industrial Robot-an International Journal, 1975

Using a Task Description Language for Assembly. The Generation of World Models

Constructing Programs with a Plan Formation System

IFAC Proceedings Volumes, Oct 1, 1983

This paper describes a robot planning system for constructing manipulator programs from task desc... more This paper describes a robot planning system for constructing manipulator programs from task descriptions. Plan formation and program generation techniques are used to transform task specifications into executable programs. The plan formation operates on a data bas e of knowledge a bo~ the world to produce the sequence of actions needed to execute the task. T~, program generation transforms the sequence of actions into an executable pro!V'am, written in MAL, using knowledge about the robot and its control syS'tem.

Detecting and forecasting economic regimes in multi-agent automated exchanges

Decision Support Systems, Nov 1, 2009

We show how an autonomous agent can use observable market conditions to characterize the microeco... more We show how an autonomous agent can use observable market conditions to characterize the microeconomic situation of the market and predict future market trends. The agent can use this information to make both tactical decisions, such as pricing, and strategic decisions, such as product mix and production planning. We develop methods to learn dominant market conditions, such as oversupply or scarcity, from historical data using Gaussian mixture models to construct price density functions. We discuss how this model can be combined with real-time observable information to identify the current dominant market condition and to forecast market changes over a planning horizon. We forecast market changes via both a Markov correction-prediction process and an exponential smoother. Empirical analysis shows that the exponential smoother yields more accurate predictions for the current and the next day (supporting tactical decisions), while the Markov correction-prediction process is better for longer term predictions (supporting strategic decisions). Our approach offers more flexibility than traditional regression based approaches, since it does not assume a fixed functional relationship between dependent and independent variables. We validate our methods by presenting experimental results in a case study, the Trading Agent Competition for Supply Chain Management.

Decision Processes in Agent-Based Contracting

The Magnet system meets many of the challenges of modeling decision making for customer agents in... more

Detecting and Forecasting Economic Regimes in Automated Exchanges

We present basic building blocks of an agent that can use observable market conditions to charact... more We present basic building blocks of an agent that can use observable market conditions to characterize the microeconomic conditions of the market and predict future market trends. The agent can use this information to make both tactical decisions such as pricing and strategic decisions such as product mix and production planning. We develop methods that can learn dominant market conditions, such as oversupply or scarcity, from historical data using computational methods to construct price density functions. We discuss how this knowledge can be used, together with real-time observable information, to identify the current dominant market condition and to forecast market changes over a planning horizon. We validate our methods by presenting experimental results in a case study, the Trading Agent Competition for Supply Chain Management.

Multi-Robot Auctions for Allocation of Tasks with Temporal Constraints

Proceedings of the AAAI Conference on Artificial Intelligence

We propose an auction algorithm to allocate tasks that have temporal constraints to cooperative r... more We propose an auction algorithm to allocate tasks that have temporal constraints to cooperative robots. Temporal constraints are expressed as time windows, within which a task must be executed. There are no restrictions on the time windows, which are allowed to overlap. Robots model their temporal constraints using a simple temporal network, enabling them to maintain consistent schedules. When bidding on a task, a robot takes into account its own current commitments and an optimization objective, which is to minimize the time of completion of the last task alone or in combination with minimizing the distance traveled. The algorithm works both when all the tasks are known upfront and when tasks arrive dynamically. We show the performance of the algorithm in simulation with different numbers of tasks and robots, and compare it with a baseline greedy algorithm and a state-of-the-art auction algorithm. Our algorithm is computationally frugal and consistently allocates more tasks than th...

Implicit Coordination in Crowded Multi-Agent Navigation

Proceedings of the AAAI Conference on Artificial Intelligence

In crowded multi-agent navigation environments, the motion of the agents is significantly constra... more In crowded multi-agent navigation environments, the motion of the agents is significantly constrained by the motion of the nearby agents. This makes planning paths very difficult and leads to inefficient global motion. To address this problem, we propose a new distributed approach to coordinate the motions of agents in crowded environments. With our approach, agents take into account the velocities and goals of their neighbors and optimize their motion accordingly and in real-time. We experimentally validate our coordination approach in a variety of scenarios and show that its performance scales to scenarios with hundreds of agents.

Monte Carlo Tree Search for Multi-Robot Task Allocation

Proceedings of the AAAI Conference on Artificial Intelligence

Multi-robot teams are useful in a variety of task allocation domains such as warehouse automation... more Multi-robot teams are useful in a variety of task allocation domains such as warehouse automation and surveillance. Robots in such domains perform tasks at given locations and specific times, and are allocated tasks to optimize given team objectives. We propose an efficient, satisficing and centralized Monte Carlo TreeSearch based algorithm exploiting branch and bound paradigm to solve the multi-robot task allocation problem with spatial, temporal and other side constraints. Unlike previous heuristics proposed for this problem, our approach offers theoretical guarantees and finds optimal solutions for some non-trivial data sets.

Adaptive Learning for Multi-Agent Navigation

Adaptive Agents and Multi-Agents Systems, May 4, 2015

When agents in a multi-robot system move, they need to adapt their paths to account for potential... more When agents in a multi-robot system move, they need to adapt their paths to account for potential collisions with other agents and with static obstacles. Existing distributed navigation methods compute motions that are optimal locally but do not account for the aggregate motions of all the agents. When there are many agents that move in a crowded space, the result is an inefficient global behavior. To address this issue, we propose a new approach which leverages techniques from machine learning and game theory. Agents using our approach dynamically adapt their motion depending on local conditions in their current environment. We validate our approach experimentally in a variety of scenarios and with different numbers of agents. When compared to other machine learning techniques, our approach produces motions that are more efficient and make better use of the space, allowing agents to reach their destinations faster.

Measuring Physiological Markers of Stress During Conversational Agent Interactions

AI for Disease Surveillance and Pandemic Intelligence, 2022

Moving in a Crowd: Safe and Efficient Navigation among Heterogeneous Agents

Multi-agent navigation methods typically assume that all agents use the same underlying framework... more Multi-agent navigation methods typically assume that all agents use the same underlying framework to navigate to their goal while avoiding colliding with each other. However, such assumption does not hold when agents do not know how other agents will move. We address this issue by proposing a Bayesian inference approach where an agent estimates the navigation model and goal of each neighbor, and uses this to compute a plan that minimizes collisions while driving it to its goal. Simulation experiments performed in many scenarios demonstrate that an agent using our approach computes safer and more time-efficient paths as compared to those generated without our inference approach and a state-of-the-art local navigation framework.

Graphically Representing Child-Robot Interaction Proxemics

This paper discusses the current analysis of a large set of child-robot interaction experiments, ... more This paper discusses the current analysis of a large set of child-robot interaction experiments, in which a 2–4 year old child plays a set of games with a small humanoid robot, with the goal of detecting early signs of autism in toddlers based on the interactions. Our first goal in this paper is to condense and display these child-robot interactions as multi-channel time series, starting with the distances between the child, robot, parent, and experimenter. Our second goal is to use these data displays to compare and contrast different children, with the aim of clustering children with similar interaction patterns. Using a ceiling-mounted camera, we record the interaction between a child and a robot which performs different games and dances. After analyzing the video footage for the locations of all people and the robot in the room over the variable length of the interaction, we condense the interactions into simplified, quantifiable, scale-invariant data. We show how the distances ...