Research
  Home | Research | Publications | Resume | Teaching | Miscellaneous | Contact

Research Statement

 
Coordination and Cooperation Among Teams of Robots, Humans and Other Agents

 

 

 

Pooyan Fazli's long term research goal is to develop highly heterogeneous teams of robots, humans, and other agents in order to work in close coordination in dynamic multi-goal environments. Robots, humans, and other agents as a team must coordinate with each otherís requirements and expectations and adapt their behaviors and decision making processes appropriately. In domains such as health care facilities, search and rescue, space exploration, and autonomous cars, effective collaboration between robots and humans will make the mission safer, faster, and more efficient and reliable. This is a relatively unexplored area of research in robotics systems.

 
 

 

Current Research

 

 

On Multi Robot Area and Boundary Coverage

 

Distributed coverage aims to deploy a team of robots to move around a target area to perform sensing, monitoring, data collection, search, or distributed servicing tasks. My Ph.D. thesis investigates three variations of the coverage problem.
First, we address the multi-robot single coverage of a target area. The aim is to guarantee that every accessible point in the area is visited in a finite time. The proposed algorithm supports heterogeneous robots having various maximal speeds, and is robust to robot failure. It also balances the workload distribution among the robots based on their maximal speeds. The obtained results on the coverage time are scalable to workspaces of different sizes, and robots of varied visual ranges.
Second, we tackle the multi-robot repeated coverage of a target area. The objective is to visit all the accessible points of the area repeatedly over time, while optimizing some performance criteria. We introduce four repeated coverage algorithms, and evaluate them under a comprehensive set of metrics including the sum of the paths/tours generated for the robots, the frequency of visiting the points in the target area, and the degree of balance in workload distribution among the robots. We also investigate the effects of environment representation, and the robots' visual range on the performance of the proposed algorithms. The results can be used as a framework for choosing an appropriate combination of repeated coverage algorithm, environment representation, and the robots' visual range based on the particular workspace and the metric to be optimized.
Third, we focus on the multi-robot repeated coverage of the boundaries of a target area and the structures inside it. Events may occur at any position on the boundaries, and the robots are not a priori aware of the event distribution. The goal is to maximize the total detection reward of the events. The reward a robot receives for detecting an event depends on how early the event is detected. To this end, we introduce an online, distributed algorithm and investigate the effects of robots' visual range, communication among the robots, and the event frequency on the performance of the algorithm.

 

 

 

 

 

 

 

Other Research Projects

 

 

Microsoft Robotics Studio Soccer Challenge

Using long-term experience of leadership and supervision of RoboCup teams in Iran, Pooyan Fazli organized the Nao Soccer Simulation team of the UBC computer science department in 2007 named "UBC Thunder". The aim of the project is to develop a dynamic task allocation mechanism in a multi-robot system using fast vision processing techniques. The team participated in the world final of robotics competitions, RoboCup 2007 (Atlanta, US), and RoboCup 2008 (Suzhou, China).

 

 

 

 

 

Unsupervised Categorization (Filtering) of Google Images based on Visual
Consistency

To furnish computers and their moving relatives, robots, with object recognition skill, a number of methods have been developed. These methods can be divided into two groups: recognition of individual objects and recognition of categories. Considerable progress has been made in the recognition of individual objects under different illumination and viewpoint conditions. Categories are more general and difficult to deal with and learning a model for them requires more complex representations. In recent years extensive studies has been carried out on particular image categories like human faces, vehicles, pedestrians, and so on. Instead, we want to develop techniques that work well to any category that a human can recognize. Also while most categorization methods require manually collecting a large sample of high quality instances of the given object category, training a classifier on them and then possibly evaluating it on more demanding new query images, our project focuses on the significant challenge of unsupervised image categorization. We have to be able to create object categories from a collection of highly noisy images with a very few or even no training examples.
The objective of this project is to study the existing methods for unsupervised object recognition and image categorization and propose a model that can learn directly from the output of image search engines (e.g. Google Images), bypassing the need to manually collect large quantities of training data. This model can then be used to refine the quality of the image search, or to search through other sources of images.

 
 

 

Past Research

 

 

Robotics Coach Simulation

Founder and Head of Robotics Coach Simulation Research Group, Department of Computer Science, Amirkabir University of Technology

Responsibility: Team Leader, Chief Designer and Developer.

The aim of the project is to design and implement an autonomous agent, which provides advice to the other agents about how to act in the environment. The team participated in the world final of robotics competitions, RoboCup 2003 (Padua, Italy).

 

 

 

 

 

 

 

 

 

 

Robotics Rescue Simulation

 

Founder and Head of Robotics Rescue Simulation Research Group, Department of Computer Science, Amirkabir University of Technology

Responsibility: Team Leader, Chief Designer and Developer.

The aim of the project is to design and develop coordination algorithms that enable teams of Ambulances, Police Forces, and Fire Brigades to save as many civilians as possible and extinguish fires in a city where an earthquake has just happened. The team participated in the world final of robotics competitions, RoboCup 2003 (Padua, Italy).

 

 
 

 

Robotics Soccer Simulation

 

Founder and Head of Robotics Soccer Simulation Research Group, Department of Computer Science, Amirkabir University of Technology.

Responsibility: Team Leader, Chief Designer and Developer.

The aim of the project is to design and implement a multi-robot system consisting of fully autonomous and heterogeneous agents to play soccer against the opponent team. The team participated in the world final of robotics competitions, RoboCup 2002 (Fukuoka, Japan), and RoboCup 2003 (Padua, Italy).

 
 

 

 

 


Copyright @ 2013 Pooyan Fazli