DynaSlum

Today, over half of the world’s population lives in urban areas and by the middle of this century 7 out of 10 people will live in a city. This increased urbanization has also lead to more and more people residing in informal settlements, generally known as slums. By the middle of the 21st century, it is estimated that the urban population of developing countries will more than double, increasing from 2.5 billion in 2009 to almost 5.2 billion in 2050. In India (2011) roughly 13.7 million households, or 17.4% of urban Indian households, were considered to be part of a slum. Each slum in a city suffers from varying degrees of depravity. This proposal aims to build high-resolution agent-based models that can describe the growth dynamics of slums in Bangalore. Such a model will create a virtual slum that decision-makers, and researchers, can use to explore how different policies would influence the growth, development or contraction of slums. The team plan to build on their existing work and develop a decision support system for slums in general, which will help guide experts when evaluating or designing policies to improve conditions within slums. This involves developing new computational methods for analyzing satellite images and new data visualization techniques for simulation steering. Through extending the existing decision support prototype, and applying it to a new domain of slum policy, the hope is to generalize the current software. Therefore, this proposal will aim to develop a reusable software framework for decision support and disseminate it to other users.

[Project Website]

The Kumbh Mela Experiment

In the Kumbh Mela Experiment we plan to develop sophisticated methods and algorithms to aid planners and event managers in managing extremely large crowds. The project aims to deliver the core components of an entire crowd management solution, all the way from designing and building personnel devices for tracking movement, to developing advanced computational models to help predict how the crowd may evolve. This leads to a vision of a living simulation where various data sources (personal devices, video cameras, smart phones, monitoring drones) can be fused into models and executed on an advanced compute infrastructure capable of providing both planning, as well as real-time reactive decision-making. This proposal addresses three key research areas related to the understanding and management of huge human crowds: Data Collection; Data Analysis; Modelling and Prediction. The Kumbh Mela Experiment will happen in April-May 2016 and will present a once in a lifetime opportunity to do both extremely exciting new science and application.

[Project website]

People

• Joris Borgdorf [eScience engineer]

SIM-CITY

By the middle of the 21st century, it is estimated that the urban population of developing counties will more than double, increasing from 2.5 billion in 2009 to almost 5.2 billion in 2050. The SIM-CITY examines urban processes within the India City of Bangalore. This includes simulation-based optimisation of fire station response, and agent-based models to understand the concept of living wage in Slums.

[Read More]

People

• Joris Borgdorf [eScience engineer]
• Harsha K [PhD student]
• Bharath Palavalli [PhD student]

TUM-CREATE

I worked as the NTU-PI of Research Package 5 (RP5) of the TUM-CREATE project. For more details of the project please visit http://tum-create.edu.sg/. I am still involved with the project as a collaborator

COSMOS – CROSIT

While I was working in Singapore I was involved in a number of projects related to Crowd Simulation. For details about all of this work please see http://crowds.sce.ntu.edu.sg

Youth Viollence

This is a project looking at developing data driven agent-based models to understand the development at identifying risk criteria for youth violence in Singapore. I am a collaborator on this project which will start officially in 2013.

MAGIC

I was also a member of the MAGIC research team at NTU: http://magic.ntu.edu.sg/

Validation through Human Computation

Agent-based modeling as a methodology for understanding natural phenomena is becoming increasingly popular in many disciplines of scientific research. Validation is still a significant problem for agent- based modelers and while various validation methodologies have been proposed, none have been widely adopted. Data plays a key role in the validation of any simulation system, typically large amounts of observable real world data are necessary to compare with model outputs. However, the complex nature of the studied natural systems will often make data collection difficult. This is certainly true for crowd and egress simulation, where data is limited and difficult to collect. In this project we attempt to develop a new technique for validation of agent-based models, particularly those which relate to human behavior. This methodology adopts ideas from the field of Human Computation as a means of collecting large amounts of contextual behavioral data. The key principle is to use games as a means of framing behavioral questions to try and capture people’s natural and instinctive decisions. We investigate some key design challenges for such games and plan to develop one example game in the form of Escape. Escape is an egress based game where people are tasked to escape from rooms inhabited by other people.

DEPATHSS

This project proposes to investigate the feasibility of using agent-based simulation and sensor technology to design an intelligent dynamic building or venue evacuation system. The system assumes an “intelligent” building, designed to have multiple static and dynamic sensors for: smoke intensity, temperature, crowd density (through image processing and CCTV), etc. During an emergency situation these sensors are used to dynamically gather information about the unfolding scenario. We consider a system that is equipped with electronic displays, which can be changed to indicate location specific directions (forward, back, left, right, down and up) by some control mechanism. With sensors and signs, it would be possible to dynamically adjust the exit directions depending on the environmental conditions and movement of the evacuees. Alone, these mechanisms (sensors and dynamic signs) would provide an effective method for a human safety controller to guide building inhabitants from the building. However, this rests on the assumption that the controller is able to make a fast and safe evacuation plan. In reality: fire spread, smoke spread, overcrowding and human panic make this planning process very complex. Moreover, the dynamic nature of the environment in emergency scenarios means events unfold in highly unpredictable ways (that would be very difficult for a fixed simulation, or prior planning to account for, e.g., bomb explosion) and effectively invalidate the effectiveness of any pre-chosen route.
DEPATHSS intends to develop a simulation system which, given sensory input from the environment, will automatically build a simulation model and then execute and evaluate a series of simulated evacuation plans to determine the safest egress plan. Once an appropriate plan is determined, the system will configure the exit signboards to direct the inhabitants to safety. The simulation will also be able to interact with the environment and effectively investigate and build its own understanding of the environment through the use of dynamic sensors.

COSMOS

11/08-11/09