From Design To Demolition

Robotics Holds a Promise to Revolutionize Every Stage Of A Building's Life Cycle

There is little doubt that the construction business is evolving through building information modeling (BIM [1] ). Having a true digital representation of a building as it is being built or renovated to compare to the design CAD allows for better design, planning and collaboration as well as avoiding costly errors. The virtualization of operations is improving efficiency to the point of being rightfully called a revolution [2] .

As soon as the proverbial cornerstone is laid, the building starts to deviate, to some degree, from the design intent captured in its digital form as a CAD model. Within generally accepted tolerances, those deviations are forgivable; however, they can still amount to important and unnecessary costs or lost revenue. For example, the consequences of even a slight dimensioning error can be eye watering when constructing luxury condominiums [3] . Furthermore, if important errors remains undetected until late in construction, the building’s structure might be compromised, making corrections very expensive.

To avoid this very problem (the late detection of the deviation from design intent or nominal values), some industries adopted metrology as a cornerstone of their production process. For example, during production, car frames and doors are routinely inspected before installation to insure a good fit, harmonizing the production chain, reducing delays, increasing productivity and normalizing costs. Borrowing from the automotive industry, the construction and renovation processes should continuously monitor and measure the building being built, updating the BIM accordingly to avoid unnecessary costs and loss of productivity.

Design and construction are, however, only the very first steps of the building's life cycle. Later stages such as operations, maintenance, renovation, repurposing, energy efficiency and even demolition and disposal could benefit just as much from BIM as the construction, especially if some degree of autonomous response is provided.

For example, once built and in operation, the building starts to deviate from its initial state : partitioning changes might be necessary to accommodate new tenants, damages caused by structural strains, water infiltrations and others will accumulate, security risks due to changes might appear and so on. All of those changes are of obvious interest to insurance companies and should be (and currently only sometimes are) thoroughly documented. Indicators of building health could also be precisely monitored: the air quality, noise levels, cleanliness, elevator inspections, presence and good functioning of security equipment (defibrillators, first aid kits, fire extinguishers) and level of consumables (from light bulbs to soap in bathrooms) are only a few examples.

Accomplishing all of those tasks require constant attention and updates to the BIM. In short, it is useful only to the degree it represents reality faithfully; otherwise, BIM might do the opposite of its intent and cause costly errors and damages rather than operational excellence. Traditionally, there are two ways of providing relevant data : fixed sensors and manual inspections.

Fixed sensors (eg. surveillance cameras, movement sensors, smoke detectors, etc.) are usually deployed in relatively small numbers in the most critical parts of the building to gather live data. One reason behind this sparseness is cost : in addition to buying standalone sensors, their installation, wiring and maintenance quickly add up to an expensive system. Moreover, complexity rapidly explodes, creating an important number of failure points. Installing a large network of fixed sensors to collect dense information throughout the building is thus unreliable and prohibitively expensive regardless whether it is wired or wireless system.

Manual inspections involve a human patrolling the area to measure with specialized equipment. The types of data that can be gathered are much more diverse than with static sensors : some examples are thermal and 3D geometry scanning, structure and elevator inspections, equipment inventory, security rounds and reading gauges that are not connected to the building numerical system. Creating a dense and complete data set is no problem either : a human can be as thorough as the task requires.

However, performing such inspections is expensive in both manpower and money. Thus, while daily manual inspections are beneficial, the cost would quickly pass beyond the profitability point of the building, leading to infrequent (at best) monitoring. Moreover, the specialized equipment used can be costly (e.g. good quality 3D scanners currently are beyond most means). For large projects, the sporadic use of such equipment (and required specialist workforce) is acceptable and profitable. However, despite having the same needs, small contractors and homeowners will find such equipment financially out of reach.

In short, both fixed sensors and manual inspections cannot provide at an economically acceptable cost the complete up-to-date information required to keep BIM truly useful throughout the life of a building. We believe that robotics will provide the answer, combining the frequency of readings from fixed sensors to the versatility and thoroughness of manual scanning all the while being cheap and simple to operate.


  1. [1] Wikipedia, Building Information Modeling.
  2. [2] Renata Barradas Gutierrez, The Digital Revolution in Construction.
  3. [3] Excelsior Measuring, About Lidar Scanning.