Project image
)}
SEK 0
pledged of SEK 39,250,000pledged of SEK 39,250,000 goal
0
backers
Funding Canceled
Funding for this project was canceled by the project creator on Jul 15 2016
SEK 0
pledged of SEK 39,250,000pledged of SEK 39,250,000 goal
0
backers
Funding Canceled
Funding for this project was canceled by the project creator on Jul 15 2016

About

I started my early studies with computer science at high school and had computers and 3D as a part of every day life like any other IT interested teenager.

I left the IT field to pursue an collage education and professional career in construction-engineering with an assigned role of project manager for a 300 M£ hospital construction project at the age of 26.

Upon learning about deep machine learning (a part of artificial intelligence) I immediately realized both the practical and economic benefits the application of deep machine learning trough computer vision and cognitive computing would bring to my field of work.

My business idea is developing a wetware (AI "software") that uses deep machine learning to "learn" the process of planing constructions just by "watching" engineers/architects working in 3D CAD-software.

Highly detailed 3DCAD-models
Highly detailed 3DCAD-models

This 3D-CAD wetware will further be helped by cognitive software that helps guide it trough the complex decisions and processes of construction design by incorporating information fetched directly from scientific texts about thermodynamics, physics, price, geometric data and so on connected to certain facts found directly in the 3D CAD-components that make up an 3D CAD-model (for example what material a 3D-component is supposed to be made of in reality). 

The softwares, tools and moments that make up the planing/design process
The softwares, tools and moments that make up the planing/design process

Step 2 in developing this intelligent construction robot will be using these 3D CAD-models to guide an industrial robot attached to a heavy terrain transport vehicle (a container stacker for example).

The industrial robot attached to the transport vehicle will be picking up the building parts from a fixed material depot on site and "stack"/"apply" the parts in to a complete building in an pre-defined order determined by the wetware.

The transport vehicle and the robot will be guided by cameras using photo-recognition to identify the individual building-parts when they are being picked up from the material depot on site. Upon picking up the correct part the vehicle and robot will be guided by following the coordinates that the wetware compiled 3D CAD-model makes up. The zero-point in this x-y-z-coordinate system would be set at the coordinates and height of one of the building corners.

incorporating topographical 3D-data in the 3D-CAD models captured with drones
incorporating topographical 3D-data in the 3D-CAD models captured with drones

 

Heavy terrain vehicles combined with industrial robots guided by computer vision aided wetware
Heavy terrain vehicles combined with industrial robots guided by computer vision aided wetware

 Buildings are what makes our entire society and around 10% of all construction costs consist of the costs for engineers and architects and around 40% of all costs consist of labor costs so there are billions of dollars in developing a wetware that just makes all the rational decisions in designing a building and further an all-terrain construction robot.

I agree with the folks at Google DeepMind I've spoken to on their estimate on this being about 5-10 years in to the future but not working out the architecture for this solution would be a shame when that day comes.

Some of the "free" benefits to this sort of construction planing are the VR-models that can be generated directly in the 3D-CAD models (seen here rendered by me in one manually compiled 3D CAD-model http://pano.autodesk.com/pano.html?url=jpgs/daf3bb83-c571-475e-9756-6be0134120f9 (open it in your phone for best effect ,and use a pair of VR-goggles)

This solution would revolutionize the entire developed part of our planet since buildings pretty much make up the physical environment of our society, It would make the work of engineers more effective if applied to interfaces like augmented reality (hololens).

Risks and challenges

Risks:

- Budget not covering all development and production costs
- Not finding the right experts in the fields of cognitive software and visual recognition

Learn about accountability on Kickstarter

Questions about this project? Check out the FAQ

Support

  1. Select this reward

    Pledge SEK 50,000 or more About US$ 5,268

    Personal Meeting

    A personal meeting in Stockholm where you get the chance to discuss the development of the product and contribute to the final result.

    Less
    Estimated delivery
    Limited 0 backers
    SEK
    Kickstarter is not a store.

    It's a way to bring creative projects to life.

    Learn more about accountability.

Funding period

- (30 days)