Cities have enormous potential to be sustainable hubs of vibrant activity. Investigate how big data, programming, machine learning, robots and 3D printing can optimise the design process and reduce resource use and waste at the same time.
VIP ChallENG research goals
Our cities face myriad challenges: climate change, rapid urbanisation and the integration of smart vehicles to name just a few. Our cities are also becoming ‘smarter’, and there are huge quantities of data – about population, transport, traffic and weather etc – up for grabs.
While the challenges increase the complexity of designing and operating our cities, this huge influx of data presents several exciting opportunities.
Throughout this course, we argue that data related to the built environment can provide valuable insights for urban planners, government agencies and the private sector to address challenges our cities face at present, and in the near and long-term future.
We will source data that have defined and will generate our cities and buildings (past, present and future) using web crawling, and will:
- Clean and prepare these data for machine learning
- Develop and program workflows that design future cities and buildings
- Feed these workflows to machines and robots to fabricate sustainable buildings
Based on various urban and architectural challenges we:
- Identify suitable data sets
- Clean data sets
- Train model with data set
- Feed generated data into individual computational design tools to generate a city or building out of data
- Order the tools to produce a design
- Fabricate buildings or building parts directly out of code
- Computer Science
- Built Environment
- Transport planning
- Data Science
- Data Analytics
- Data Visualisation
- Manufacturing Engineering
Explore the UrbanAI Projects
Forming agile teams and using the scrum method to work together, we will develop and work on different projects. The overall agenda is to provide the foundation for a synthetic design method that combines machine learning and computational design to design sustainable and liveable cities. Below are the various aspects you can choose to explore.
This project aims to collect as many data suitable for the build environment as possible using: web and HTML scraping, storing of data in REST framework, cleaning and preparing of sourced data sets for later machine learning.
This project aims to build a 3D design framework as 'spine' or 'conveyor belt' which allows: a) data sourced in Project 1. to feed in; b) data gained through machine learning in Project 4. to feed in; c) sequentially organised computational tasks in Product 3. to feed in.
This project uses task planning to establish a sequential order to tasks using ML. Tasks are computational design scripts in either Grasshopper, Python with inputs and/or outputs that need to be ordered by ML.
This project aims to develop machine learning tools used for urban design in Western Sydney and architecture in designing a future entry to the Solar Decathlon sustainable housing competition.
A robot cannot use a 2D drawing but needs code to run. Using programs that can generate design and feed them into the machine, and using machine constraints to guide the design, are central to this project.