Home research People General Info Seminars Resources Intranet
| Algorithms & Applcations Group | Home | Research | Publications | People | Resources | News
Home Page for Marco Morales | Parasol Laboratory


Picture Marco Morales
Graduated, now Assistant Professor at ITAM
Algorithms & Applications Group

Parasol Laboratory url: http://parasol.tamu.edu/~marcom/
Department of Computer Science and Engineering email:
Texas A&M University office: 407C HRBB
College Station, TX 77843-3112 tel: (979) 847-8835
USA fax: (979) 458-0718


Curriculum Vitae Statements: Research, Teaching Publications Courses and Miscellanea


Marco Morales is an Assistant Professor at Instituto Tecnológico Autónomo de México since August 2007. Please visit Marco Morales' ITAM web page .

He obtained a Ph.D. in Computer Science from the Department of Computer Science of Texas A&M University where he worked under the supervision of Dr. Nancy M. Amato in the Algorithms & Applications Group of the Parasol Lab. He got a B.S. in Computer Engineering and a M.S. in Electrical Engineering, both from the Facultad de Ingeniería of the Universidad Nacional Autónoma de México (UNAM), and later he was awarded a Fulbright/García-Robles scholarship to pursue his PhD.

Morales' research is focused in motion planning and supporting areas, such as machine learning and computational geometry, and its applications to areas such as robotics, bioinformatics, and computational neuroscience. In addition, his interests include artificial intelligence, optimization, computational biology, architectures, parallel computing, and interfaces.


Algorithmic Techniques for Probabilistic Motion Planning:

The goal of Motion Planning is to find valid paths for a movable object (robot) between pairs of configurations. This intractable problem is addressed with probabilistic motion planners that model the connectivity of the robot's configuration space through random sampling of valid configurations and transitions between configurations.

Feature Sensitive Motion Planning
None of the many motion planners available performs optimally in every case. Rather, performance depends on the strengths and weaknesses of each planner. Also, the planning space may have vastly different regions, each better suited for a particular planner.
We propose a feature-sensitive meta-planner that cooperatively applies different planners to map the planning space. Features are computed to characterize the planning space in order to find an appropriate partitioning and to assign the best planning strategy for each region according to the features of both regions and planners. Regional maps are combined into a global map that represents the overall problem.
We show how this feature-sensitive strategy produces cheaper and better roadmaps than those obtained with simple planners only.
The contribution of this work is the mechanism to collect information about the sampling process to understand the complexity of the space and automatically adjust and focus planners.

Metrics for Probabilistic Motion Planning
What is the best planner for a problem?, How does the mapping effectiveness of a planner change over time?, Can we find the best mapping strategy for certain areas of the planning space?. In order to address these challenging questions we need to characterize the planning process.
We propose a set of metrics that measure how each new sample improves, or not, a C-space model. We show how, using only local information, this characterization enables analyses of how different planning strategies explore the configuration space.
We have already applied these metrics to determine when a particular planning strategy has 'converged' on the best C-space model that it is capable of building and to analyze the roadmaps obtained with different planning strategies.

Improving the Connectivity of PRM Roadmaps
We investigate how to improve the coverage and connectedness of roadmaps produced with probabilistic roadmap methods (PRMs) by expanding the PRM framework with a step to attempt connections between existing connected components.
We coordinate the application of a suite of powerful connection methods, such as Rapidly-exploring Randomized Trees (RRTs) and a ray-tracing based method. A scheduler selects pairs of nodes from different components and assigns the order of connection attempts based on multiple criteria such as distance between components and node density.
This simple method to identify important or promising regions for exploration also provides a mechanism for controlling the cost of the connection attempts. We obtained significant roadmap improvements in times that are on the same order as the times used to generate the initial roadmap.

Iterative Map Generation (IMG)
One important practical issue in randomized planning is the lack of automated mechanisms to determine how large a roadmap is needed for a given problem.
Incremental Map Generation is a new framework based on probabilistic roadmap methods that iteratively adds increment of samples and connections to the evolving roadmap until it stops improving.
A set of evaluation criteria determines whether the planning strategy is being effective at improving the roadmap. We propose some general evaluation criteria and show how to apply them to construct different types of roadmaps, e.g., roadmaps that coarsely or more finely map the space. In addition, we show how IMG can be integrated with previously proposed adaptive strategies for selecting sampling methods. We provide results illustrating the power of IMG.

Biasing and Composing Motion Planners
With the success of probabilistic planners, much work has been done to design new sampling techniques and distributions. However, little work has been done to combine these techniques to create new sampling distributions.
We propose to bias a sampling distribution with another such that the resulting composed distribution out-performs either of its parent distributions. We present a general framework for biasing samplers that is easily extendable to new distributions and can handle an arbitrary number of parent distributions by chaining them together.
Our experimental results show that by combining distributions, we can out-perform existing planners. We also notice that no single combination of distributions performs the best in all problems, and we identify which combinations perform better for the specific application domains studied.

User-Guided Motion Planning
Although randomized motion planners perform well on most cases, there still exist some problems that are beyond their capabilities due to computational or time limitations. There are situations where automatic methods fail to discover some `critical' configurations that are naturally apparent to a human observer.
We consider how to incorporate the strengths of both human operators and automatic planning methods with the help of customized interfaces.

Applications for Computational Neuroscience:

Neuron PRM: A Framework for Constructing Cortical Networks
The brain has extraordinary computational power to represent and interpret complex natural environments. These natural computations are essentially determined by the topology and geometry of the brain's architectures.
We present a framework to construct a 3D model of a cortical network using a probabilistic approach inspired by the motion planning probabilistic roadmap methods (PRMs). We represent the network as a large-scale directed graph, and use L-systems and statistics data to `grow' neurons that are morphologically indistinguishable from real neurons.
Although not the usual motion planning problem, our objective of building a network that encodes the pathways of the cortical network is analogous to the objective of PRMs of constructing roadmaps that contain a representative sample of feasible paths.

Interfaces:

Vizmo++: A Visualization/Authoring Tool for Motion Planning
VIZMO is a 3D visualization/authoring tool for files provided/generated by OBPRM motion planning library. This new version of VIZMO, VIZMO++, was developed for visualizing and editing motion planning environments, problem instances, and their solutions. VIZMO++ offers a nice and easy to use graphical user interface (GUI) that allows you to display workspace environments, roadmap, path, and start/goal positions. It enables users to interact with and edit the environment. For example, it lets users manipulate obstacles and robot configurations, set queries, save new environments to be able to work on them later, or select and move nodes and thus editing existing or creating new paths and roadmaps. VIZMO++ interfaces with the planners available in our motion planning library to enable users to run new queries. Our tool provides a convenient interface to select planners and set their parameters.


Publications

Planning Motions for Shape-Memory Alloy Sheets, Daniel Tomkins, Mukulika Ghosh, Jory Denny, Samuel Rodriguez, Marco Morales, Nancy M. Amato, Technical Report, TR15-001, Parasol Lab, Texas A&M University, College Station, TX, Jan 2015.
Technical Report(pdf, abstract)

Adapting RRT Growth for Heterogeneous Environments, Jory Denny, Marco A. Morales A., Samuel Rodriguez, Nancy M. Amato, In Proc. IEEE Int. Conf. Intel. Rob. Syst. (IROS), pp. 1772 - 1778, Tokyo, Japan, Nov 2013.
Proceedings(ps, pdf, abstract)

Structural Improvement Filtering Strategy for PRM, Roger Pearce, Marco Morales, Nancy M. Amato, In Proc. Int. Conf. on Robotics: Science and Systems, pp. 167-174, Zurich, Switzerland, Jun 2008.
Proceedings(pdf, abstract)

Metrics for Sampling-Based Motion Planning, Marco Morales, Ph.D. Thesis, Parasol Laboratory, Department of Computer Science, Texas A&M University, College Station, Texas, Dec 2007.
Ph.D. Thesis(pdf, abstract)

Biasing Samplers to Improve Motion Planning Performance, Shawna Thomas, Marco Morales, Xinyu Tang, Nancy M. Amato, In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pp. 1625-1630, Rome, Italy, Apr 2007.
Proceedings(ps, pdf, abstract)

Incremental Map Generation (IMG), Dawen Xie, Marco Morales, Roger Pearce, Shawna Thomas, Jyh-Ming Lien, Nancy M. Amato, In Proc. Int. Wkshp. on Alg. Found. of Rob. (WAFR), New York City, NY, Jul 2006. Also, Technical Report, TR06-005, Department of Computer Science and Engineering, Texas A&M University, Mar 2006.
Proceedings(ps, pdf, abstract) Technical Report(ps, pdf, abstract)

Metrics for Comparing C-space Roadmaps, Marco A. Morales A., Roger Pearce, Aimée Vargas E., Nancy M. Amato, Technical Report, TR05-012, Parasol Laboratory, Department of Computer Science, Texas A&M University, College Station, Texas, U.S.A., Sep 2005.
Technical Report(ps, pdf, abstract)

C-Space Subdivision and Integration in Feature-Sensitive Motion Planning, Marco A. Morales A., Lydia Tapia, Roger Pearce, Samuel Rodriguez, Nancy M. Amato, In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pp. 3114-3119, Barcelona, Spain, May 2005. Also, Technical Report, TR04-004, Parasol Laboratory, Department of Computer Science, Texas A&M University, College Station, Texas, U.S.A., Sep 2004.
Proceedings(ps, pdf, abstract) Technical Report(ps, pdf, abstract)

A Machine Learning Approach for Feature-Sensitive Motion Planning, Marco Morales, Lydia Tapia, Roger Pearce, Samuel Rodriguez, Nancy M. Amato, In Proc. Int. Wkshp. on Alg. Found. of Rob. (WAFR), pp. 361-376, Utrecht/Zeist, The Netherlands, Jul 2004. Also, Technical Report, TR04-001, Parasol Laboratory, Department of Computer Science, Texas A&M University, College Station, Texas, U.S.A., Feb 2004.
Proceedings(ps, pdf, abstract) Technical Report(ps, pdf, abstract)

Improving the Connectivitiy of PRM Roadmaps, Marco Morales, Samuel Rodriguez, Nancy M. Amato, In Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pp. 4427-4432, Taipei, Taiwan, Sep 2003.
Proceedings(ps, abstract)

Neuron PRM: A Framework for Constructing Cortical Networks, Jyh-Ming Lien, Marco Morales, Nancy M. Amato, Neurocomputing, 52-54(28):191-197, Jun 2003. Also, Technical Report, TR01-002, Parasol Laboratory, Department of Computer Science, Texas A&M University, Oct 2001.
Journal(ps, pdf, abstract) Technical Report(ps, pdf, abstract)