Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Machine Learning use in
Construction
All rights reserved. AI.Business Owner: http://ai.business
USE CASE – ATMA: Autonomous TMA
Truck
All rights reserved. AI.Business Owner: http://ai.business
The automated vehicle “le...
ATMA (Autonomous TMA Truck):
EFFECTS OF USAGE
• Is outfitted with an electro-mechanical system and fully integrated
sensor...
USE CASE – LENS: Predictive modeling
All rights reserved. AI.Business Owner: http://ai.business
Lens is a model-based esti...
LENS - Predictive modeling:
EFFECTS OF USAGE
• The platform allows users to work with any model, regardless of
modeling st...
USE CASE – Raising Efficiency of Air
Conditioning Systems in Commercial Buildings
All rights reserved. AI.Business Owner: ...
Raising Efficiency of Air Conditioning Systems in
Commercial Buildings: EFFECTS OF USAGE
• The application models the effe...
USE CASE – Detecting Building Collapse in
Post-Earthquake Environments
All rights reserved. AI.Business Owner: http://ai.b...
Detecting Building Collapse in Post-Earthquake
Environments: EFFECTS OF USAGE
• For earthquakes in particular, being able ...
USE CASE – Earthquake-Induced
Structural Damage Classifier
All rights reserved. AI.Business Owner: http://ai.business
Deve...
Earthquake-Induced Structural
Damage Classifier: EFFECTS OF USAGE
• Used for predicting the post-earthquake damage state, ...
USE CASE – Fatigue Crack Sensor
All rights reserved. AI.Business Owner: http://ai.business
Fatigue Crack Sensor:
EFFECTS OF USAGE
• The machine learning platform uses different input feature combinations based on
...
USE CASE – Mapillary: City planning
and inventory of roads and signage
All rights reserved. AI.Business Owner: http://ai.b...
Mapillary - City planning and inventory of roads
and signage: EFFECTS OF USAGE
• With Mapillary photos included in the str...
USE CASE – Smart buildings
All rights reserved. AI.Business Owner: http://ai.business
Data-enabled machine learning create...
Smart buildings: EFFECTS OF USAGE
• Smart building technology learns and anticipates the user's preferences, and altering
...
USE CASE – Estimating energy
performance of residential buildings
All rights reserved. AI.Business Owner: http://ai.busine...
Estimating energy performance of
residential buildings: EFFECTS OF USAGE
• Extensive simulations on 768 diverse residentia...
USE CASE – Improving BEMS (Building
Energy Management)
All rights reserved. AI.Business Owner: http://ai.business
The 3rd ...
Improving BEMS (Building Energy
Management): EFFECTS OF USAGE
• The system combines an energy model of the building with e...
Upcoming SlideShare
Loading in …5
×

10 uses cases - Artificial Intelligence and Machine Learning in Construction - by ai.business

10 uses cases - Artificial Intelligence and Machine Learning in Construction
Created by http://ai.business

  • Login to see the comments

10 uses cases - Artificial Intelligence and Machine Learning in Construction - by ai.business

  1. 1. Machine Learning use in Construction All rights reserved. AI.Business Owner: http://ai.business
  2. 2. USE CASE – ATMA: Autonomous TMA Truck All rights reserved. AI.Business Owner: http://ai.business The automated vehicle “learns” from the human-driven one that is outfitted with a NAV Module that is strapped to the roof of the vehicle during testing. This transmits the GPS position data called "eCrumbs" back to the Follower vehicle, which then uses the data to follow the exact path and speed of the Leader vehicle at each point along the route.
  3. 3. ATMA (Autonomous TMA Truck): EFFECTS OF USAGE • Is outfitted with an electro-mechanical system and fully integrated sensor suite, enabling Leader/Follower capability; • This system configuration permits replicating the real-world operation; • ATMA can follow a lead vehicle completely unmanned; • The NAV Module can be easily unstrapped and removed from one vehicle and installed on another if a different leader vehicle is required. Source: http://www.royaltruckandequipment.com/atma All rights reserved. AI.Business Owner: http://ai.business
  4. 4. USE CASE – LENS: Predictive modeling All rights reserved. AI.Business Owner: http://ai.business Lens is a model-based estimating tool that ties the Autodesk Revit Building Information Model (BIM) to Art to Science Estimating (ASE) at the earliest stages of a project.
  5. 5. LENS - Predictive modeling: EFFECTS OF USAGE • The platform allows users to work with any model, regardless of modeling standards, at the earliest stages. • Lens improves the speed in which a takeoff can be competed and the reduction in time required to update estimates makes the entire preconstruction. • It helps identify items that may be normally missed due to the possibility to see the intent of the designer as early as design development. Source: http://www.jedunn.com/blog/what-lens All rights reserved. AI.Business Owner: http://ai.business
  6. 6. USE CASE – Raising Efficiency of Air Conditioning Systems in Commercial Buildings All rights reserved. AI.Business Owner: http://ai.business Machine learning techniques can be used to predict building A/C energy consumption to help with efficiently automating the air conditioning process.
  7. 7. Raising Efficiency of Air Conditioning Systems in Commercial Buildings: EFFECTS OF USAGE • The application models the effect of each building sensor measurement on the A/C system energy consumption • 3rd order polynomial support vector regression (SVR) model best predicts the building A/C system • Uses supervised learning algorithms to predict the amount of energy consumed to maintain the temperature at a desirable level. • Artificial neural networks achieve good results in predicting consumed energy in commercial buildings and offices. Source: http://cs229.stanford.edu/proj2013/MahdiehMohammadiEhsani- Y2E2BuildingEnergyStudy.pdf All rights reserved. AI.Business Owner: http://ai.business
  8. 8. USE CASE – Detecting Building Collapse in Post-Earthquake Environments All rights reserved. AI.Business Owner: http://ai.business
  9. 9. Detecting Building Collapse in Post-Earthquake Environments: EFFECTS OF USAGE • For earthquakes in particular, being able to map the distribution of damage quickly and with confidence can help channel appropriate aid to the most- severely impacted regions. • Accurate mapping can also aid in determining whether citizens can return safely to their homes, so as to prevent casualties from delayed building collapses. • Using the machine learning techniques developed, future disaster relief professionals might be able to use a more limited field-based damage assessment, in combination with remote-sensing-based data, to identify highly damaged areas more quickly and at lower cost. Source: http://cs229.stanford.edu/proj2013/BenjaminCorriganGibbsWong- UsingLowCostRemoteSensingDataToDetectBuildingCollapseInPostEarthquakeEnvironments.pdf All rights reserved. AI.Business Owner: http://ai.business
  10. 10. USE CASE – Earthquake-Induced Structural Damage Classifier All rights reserved. AI.Business Owner: http://ai.business Developing a structural damage classifier using support vector machines.
  11. 11. Earthquake-Induced Structural Damage Classifier: EFFECTS OF USAGE • Used for predicting the post-earthquake damage state, given the building features and input ground motion. • Used for accelerating post-earthquake damage evaluation of critical buildings. This will allow faster recovery time and decrease financial losses expected from downtime and repair. • Using k-means clustering, each ground motion is categorized based on frequency content. • The most influential feature is the correlation between the fundamental period and the earthquake type. • A preliminary safety evaluation of a building is possible using this damage state classifier. Source: http://cs229.stanford.edu/proj2015/343_poster.pdf All rights reserved. AI.Business Owner: http://ai.business
  12. 12. USE CASE – Fatigue Crack Sensor All rights reserved. AI.Business Owner: http://ai.business
  13. 13. Fatigue Crack Sensor: EFFECTS OF USAGE • The machine learning platform uses different input feature combinations based on sensor data that are defined and tested, and different classification methods are utilized to determine a specimen is intact or damaged. • The sensor data is acquired from steel specimen using a high-frequency fatigue crack sensor. • The raw sensor data is pre-processed so that several features representing meaningful information of sensor data can be extracted. Source: http://cs229.stanford.edu/proj2015/341_poster.pdf All rights reserved. AI.Business Owner: http://ai.business
  14. 14. USE CASE – Mapillary: City planning and inventory of roads and signage All rights reserved. AI.Business Owner: http://ai.business Mapilliary uses machine learning to stitch together 3D visualizations of photos contributed by its more than 12,000 users. The images are available via an API.
  15. 15. Mapillary - City planning and inventory of roads and signage: EFFECTS OF USAGE • With Mapillary photos included in the strategic planning process of a city employees from all departments in the municipality will be able to see their future investment areas combined with up to date photos on the map. • The Mapillary mobile application can be used along a selected railway line to complete field observations and quality controls. • Main advantages: – Turn street photos into 3D maps within minutes – View, edit, and extract geospatial data – Automate hours of manual work with one click. Source: 1) https://www.mapillary.com/ ; 2) http://www.informationweek.com/strategic-cio/executive-insights-and-innovation/11- cool-ways-to-use-machine-learning/d/d-id/1323375?image_number=13 All rights reserved. AI.Business Owner: http://ai.business
  16. 16. USE CASE – Smart buildings All rights reserved. AI.Business Owner: http://ai.business Data-enabled machine learning creates a smart building, whose defining feature is the ability to be proactive in making appropriate changes to services on behalf of its users. Smart = “equal to selfawareness plus the ability to react”. (Andrew Eastwell, Chief Executive of the Building Services Research and Information Association)
  17. 17. Smart buildings: EFFECTS OF USAGE • Smart building technology learns and anticipates the user's preferences, and altering conditions to meet his needs more precisely and exactly than we ourselves can. • The huge amount of data can help the smart buildings to make reactive – and even anticipatory and personalized – real-time alterations to a building’s environment to suit its occupants. • The same technology can prolongue the time that elderly people can remain in their own homes by allowing remote monitoring of health through blood pressure and heart monitors that note behavior patterns and highlight any change that might indicate a problem. • Smart Buildings are conceived as upgradeable due to the fact that technology changes, with elements added in such a way that they can easily be changed as technology and the building’s use develop. Source: http://www.raeng.org.uk/publications/reports/raeng-smart-buildings-people-and- performance All rights reserved. AI.Business Owner: http://ai.business
  18. 18. USE CASE – Estimating energy performance of residential buildings All rights reserved. AI.Business Owner: http://ai.business Developing a statistical machine learning framework to study the effect of eight input variables (relative compactness, surface area, wall area, roof area, overall height, orientation, glazing area, glazing area distribution) on two output variables, namely heating load (HL) and cooling load (CL), of residential buildings.
  19. 19. Estimating energy performance of residential buildings: EFFECTS OF USAGE • Extensive simulations on 768 diverse residential buildings show that we can predict HL and CL with low mean absolute error deviations from the ground truth which is established using Ecotect (0.51 and 1.42, respectively). • The results support the feasibility of using machine learning tools to estimate building parameters as a convenient and accurate approach, as long as the requested query bears resemblance to the data actually used to train the mathematical model in the first place. Source: http://www.sciencedirect.com/science/article/pii/S037877881200151X All rights reserved. AI.Business Owner: http://ai.business
  20. 20. USE CASE – Improving BEMS (Building Energy Management) All rights reserved. AI.Business Owner: http://ai.business The 3rd wave of innovation will take this analysis concept even further to “optimization”. (Mike Zimmerman, Founder of BuildingIQ)
  21. 21. Improving BEMS (Building Energy Management): EFFECTS OF USAGE • The system combines an energy model of the building with external data such as weather forecasts and energy pricing signals to automatically write set points for the BEMS and execute Demand Response (DR) events. • The SaaS (cloud based) software works with the buildings existing BEMS and utility demand response systems, incorporating weather forecasts, occupant comfort, utility prices and demand response signals into its optimization algorithms. Source: http://www.memoori.com/the-future-impact-of-machine-learning-predictive-analysis-on- building-energy-management/ All rights reserved. AI.Business Owner: http://ai.business

×