While autonomous cars are on the front pages of every publication, there is a silent revolution underway in the industrial counterparts of passenger cars. In mining industry, fully automated hauling trucks have been in production for years. For example, Rio Tinto's fleet of driverless trucks currently hauls 25% of the ore and waste in Australia's Pilbara region. They are using a fleet of 80 autonomous trucks from Komatsu. They operate many hours longer and at 15% or lower cost compared to manually operated trucks.
Momentum is now picking up in the construction industry. Here we discuss some of the recent developments related to construction robotics and what to expect going forward.
While most of the world is concerned about loss of jobs due to robotics, construction industry has the opposite problem. The workforce is getting older. Fewer and fewer millennials want to be employed as construction workers. As a result, the industry is begging for autonomous robots which can help existing workers become more productive.
Intelligent Machine Control
Komatsu introduced fully automatic blade control with their D61i-23 dozer in 2013 along with their Intelligent Machine Control technology. The machine supports manual and autonomous modes for dozing. Exact positioning involves three components - a cab top GNSS antenna for satellite based positioning, en enhanced inertial measurement unit to calculate positional changes to cover for low GPS resolution, and stroke-sensing hydraulic cylinders to measure blade angle and tilt accurately. Once the job is specified using the software, the blade can operate autonomously. It even detects when too much load can cause the vehicle to slip from the track and compensate by adjusting blade elevation. When closer to finish, the blade adjusts accordingly for finish grade. Overall, this reduces the manual effort needed to control the blade allowing faster, more efficient operations.
Royal Truck & Equipment partnered with Micro Systems, Inc to allow a fleet of trucks to navigate together. Trucks follow a lead vehicle that transmits GPS data that instructs others when to turn, brake and accelerate/decelerate.
Digitization in Construction
More and more startups and large companies are competing to digitize construction workflow. Companies such as PlanGrid are revolutionizing this area. As digital files become more and more common, this further incentivizes robotic automation, which depends on 3D CAD models and design files.
So far, automation has been focused on enhancing the ability of manual operators through automatic blade control and other technologies. Complete automation has only been possible in large mining sites where the terrain is well understood and there is less need for collaboration with human workers.
Completely unmanned operations in smaller, busier construction sites while collaborating with workers will be the next goal. Safety is a key concern here. With workers operating in the same environment and sharing space, fully unmanned operations will require highly advanced safety features to avoid any incidents. This technology is somewhat related to what autonomous cars need, but there are large differences. For one, the level of dust and other hindrances to visibility is higher. Second, there are no clearly marked lanes and ability to create HD maps ahead of time. Third, unlike cars, these equipment are expected to manipulate the ground and environment ( instead of simply navigating ) making the challenges different from cars simply passing on roads.
Other than horizontal construction, robotics is transforming vertical construction as well. Brick-laying robots are already available in the market. New robots will be able to perform mundane tasks such as drilling holes in dry walls for wiring.
If you are seeking partners to accelerate your automation needs related to heavy equipment including autonomous sensing and operations, write to us at firstname.lastname@example.org. We develop computer vision algorithms for regular camera as well as 3D cameras/LiDAR. We also provide synthetic training data to accelerate your computer vision pipeline.