Abstractions for Robots Interacting with Open Worlds
Robots that are capable of successfully operating in open environments, often interacting with other robots and humans, must find ways to cope with uncertainty and incompleteness of information across a hierarchy of scales ranging from the lower levels of sensorimotor signals to more temporally extended aspects of autonomous behaviour. I will present results from recent work in my group that address two aspects of this broad challenge - defining and learning action-oriented symbols; and using categorisation over a space of policies to achieve flexible interaction with other agents.
I will first set the scene by mentioning a few examples of generative models that can be used for intention, activity and choice prediction, as applied to robotic systems. I will argue that success of these models depends on having good hierarchical representations.
Motivated thus, I will describe a novel algorithm for multiscale classification of trajectories in robotic systems. Unlike previous sampling-based approaches in robotics, which use graphs to capture information about the path-connectedness of a configuration space, we construct a multiscale approximation of neighborhoods of the collision free configurations based on filtrations of simplicial complexes. Our approach thereby extracts additional homological information, which is essential for a topological trajectory classification of sets of trajectories starting and ending in two fixed points. Using a cone construction, we then further generalize this approach to classify sets of trajectories even when start and end points are allowed to vary in a path-connected subset. We furthermore show how an augmented filtration of simplicial complexes based on an arbitrary function on the configuration space, such as a costmap, can be defined to incorporate additional constraints. We evaluate this in up to 6-dim configuration spaces, in simulation as well as real world experiments with the Baxter and PR2 robots.
Next, I will outline a model for ad hoc multi-agent interaction without prior coordination, which utilises categorisation in an explicitly strategic setting. By conceptualizing the interaction as a stochastic Bayesian game, the choice problem is formulated in terms of types in an incomplete information game, allowing for a learning algorithm that combines the benefits of Harsanyi’s notion of types and Bellman’s notion of optimality in sequential decisions. I will quote some results pertaining to the optimality of this algorithm even when it is based on a type space that is incorrect in a sense. I will also present preliminary results from experiments involving human-machine interaction where we show a better rate of coordination than alternate multi-agent learning algorithms.
Time permitting, I will conclude with a brief description of some problem domains we are pursuing in our robotics laboratory, that demand such a hierarchical treatment of uncertainty.Robots that are capable of successfully operating in open environments, often interacting with other robots and humans, must find ways to cope with uncertainty and incompleteness of information across a hierarchy of scales ranging from the lower levels of sensorimotor signals to more temporally extended aspects of autonomous behaviour. I will present results from recent work in my group that address two aspects of this broad challenge - defining and learning action-oriented symbols; and using categorisation over a space of policies to achieve flexible interaction with other agents.
Dr. Subramanian Ramamoorthy is a Reader (Associate Professor) in the School of Informatics, University of Edinburgh, where he has been since 2007. He is Coordinator of the EPSRC Robotarium Research Facility, and Executive Committee Member for the Centre for Doctoral Training in Robotics and Autonomous Systems. Previously, he received a PhD in Electrical and Computer Engineering from The University of Texas at Austin. He is a Member of the Young Academy of Scotland at the Royal Society of Edinburgh, where he co-chairs the Industry Working Group.
His current research is focussed on problems of autonomous learning and decision-making under uncertainty, by long-lived agents and agent teams interacting within dynamic environments. This work is motivated by applications in autonomous robotics, human-robot interaction, intelligent interfaces and other autonomous agents in mixed human-machine environments. These problems are solved using a combination of methods involving layered representations based on geometric/topological abstractions, game theoretic and behavioural models of inter-dependent decision making, and machine learning with emphasis on issues of transfer, online and reinforcement learning.
His work has been recognised by nominations for Best Paper Awards at major international conferences - ICRA 2008, IROS 2010, ICDL 2012 and EACL 2014. He serves in editorial and programme committee roles for conferences and journals in the areas of AI and Robotics. He leads Team Edinferno, the first UK entry in the Standard Platform League at the RoboCup International Competition. This work has received media coverage, including by BBC News and The Telegraph, and has resulted in many public engagement activities, such as at the Royal Society Summer Science Exhibition, Edinburgh International Science festival and Edinburgh Festival Fringe.
Before joining the School of Informatics, he was a Staff Engineer with National Instruments Corp., where he contributed to five products in the areas of motion control, computer vision and dynamic simulation. This work resulted in seven US patents and numerous industry awards for product innovation.