Applied Motion Lab

Our research focuses on using computers to understand, recreate, and interact with the dynamic nature of the world. Current research projects include Crowd Simulation, Virtual Humans, and Robot Motion Planning.

Watch Videos
Data Driven Sokoban Puzzle Generation with MCTS (image missing)
AIIDE 2016 Full Paper

Data Driven Sokoban Puzzle Generation with MCTS

In this work, we propose a Monte Carlo Tree Search (MCTS) based approach to procedurally generate Sokoban puzzles. Our method generates puzzles through simulated game play, guaranteeing solvability...

Full Report (PDF)
Implicit Coordination in Crowded Multi-Agent Navigation (image missing)
AAAI 2016 Full Paper

Implicit Coordination in Crowded Multi-Agent Navigation

In crowded environments, agent paths are significantly constrained by nearby neighbors, leading to difficulty in generating globally efficient motion. To address this, we propose a new distributed approach to coordinate the motions of agents in crowded environments.

Full Report (PDF)Watch Video
Velocity-Based Modeling of Physical Interactions in Dense Crowds (image missing)
The Visual Computer, 2015 Article

Velocity-Based Modeling of Physical Interactions in Dense Crowds

We present an algorithm to simulate physics-based interactions in dense crowds. Our approach models both physical forces and interactions between agents and obstacles, while also allowing agents to anticipate and avoid upcoming collisions during navigation.

Full Report (PDF)Watch Video
Prioritized Group Navigation with Formation Velocity Obstacles (image missing)
ICRA 2015 Full Paper

Prioritized Group Navigation with Formation Velocity Obstacles

We introduce the problem of navigating a group of robots having prioritized formations amidst static and dynamic obstacles. Our formulation allows users to define a number of template formations, each with a specified priority value...

Full Report (PDF)Watch Video
Stochastic tree search with useful cycles for patrolling problems (image missing)
ICRA 2015

Stochastic tree search with useful cycles for patrolling problems

An autonomous robot team can be employed forcontinuous and strategic coverage of arbitrary environmentsfor different missions. In this work, we propose an anytimeapproach for creating multi-robot patrolling policies. Ourapproach involves a novel extension of Monte Carlo TreeSearch (MCTS) to allow robots to have life-long, cyclic policiesso as to provide continual coverage of an environment. Ourproposed method can generate near-optimal policies for a teamof robots for small environments in real-time (and in largerenvironments in under a minute). By incorporating additionalplanning heuristics we are able to plan coordinated patrollingpaths for teams of several robots in large environments quicklyon commodity hardware.

Full Report (PDF)Watch Video
Anytime Navigation with Progressive Hindsight Optimization (image missing)
IROS 2014 Full Paper

Anytime Navigation with Progressive Hindsight Optimization

In multi-robot systems, efficiently navigating in a partially-known environment is an ubiquitous but challenging task, as each robot must account for the uncertainty introduced, for example, by other moving robots. This uncertainty makes pre-computed plans not always applicable..

Full Report (PDF)Watch Video
Adaptive Learning for Multi-Agent Navigation (image missing)
AAMAS 2015 Full Paper

Adaptive Learning for Multi-Agent Navigation

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..

Full Report (PDF)Watch Video
User-driven narrative variation in large story domains using monte carlo tree search (image missing)
AAMAS 2014

User-driven narrative variation in large story domains using monte carlo tree search

Planning-based techniques are powerful tools for automatednarrative generation, however, as the planning domain growsin the number of possible actions traditional planning techniques su er from a combinatorial explosion. In this work, we apply Monte Carlo Tree Search to goal-driven narrative generation. We demonstrate our approach to have an order of magnitude improvement in performance over traditional search techniques when planning over large story domains. Additionally, we propose a Bayesian story evaluationmethod to guide the planning towards believable narrativeswhich achieve user-de ned goals. Finally, we present an interactive user interface which enables users of our frameworkto modify the believability of different actions, resulting in greater narrative variety.

Full Report (PDF)Watch Video
Automated RPG Map Generation (image missing)
Honor's Thesis (Nate Buck, 2013)

Automated RPG Map Generation

This world generation system for a turn-based strategy/role-playing game, is implemented using a combination of a fractal-based algorithm and a Markov chain matrix process. The process includes generating a political map, path-finding along important cities, and generating discrete levels along determined stochastically...

Full Report (PDF)
Automated RPG Map Generation (image missing)
MiG 2013, Conference Paper

Automated RPG Map Generation

This world generation system for a turn-based strategy/role-playing game, is implemented using a combination of a fractal-based algorithm and a Markov chain matrix process. The process includes generating a political map, path-finding along important cities, and generating discrete levels along determined stochastically...

Full Report (PDF)
C-OPT: Coverage-Aware Trajectory Optimization Under Uncertainty (image missing)
RAL 2016

C-OPT: Coverage-Aware Trajectory Optimization Under Uncertainty

In this paper we introduce a new problem of continuous, coverage-aware trajectory optimization under localization and sensing uncertainty

Full Report (PDF)Watch Video

Multi-Agent Navigation

Work on the theoretical problem of getting agents to their goals in the face of noise, uncertainty, and the motion of other agents, with applications in robotics and AI.

Physically-Based Animation

This work attempts to recreate real-world phenomena through the virtual application of physical laws.

Games and VR

Various work in improving the state-of-the-art in computer games and virtual reality.