The development of autonomous robotic vehicles (AV’s) for construction seems to make great sense. Potential productivity gains, a solution to the shortage of new entrants to the skilled labour force, and a desire to improve safety conditions for site workers make robotic vehicles an attractive, efficient alternative to manual operation of earth-moving equipment and vehicles.
The path towards full AV construction has five identifiable steps:
- Full manual, where the operator sitting in the machine operates all functions;
- In-sight tele-operation, where a nearby operator uses a remote control;
- Tele-remote operation, where the operator is in a control room receiving audio and video feedback;
- Semi-autonomous operations, where programmed repetitive functions are simply overseen by the operator;
- Full autonomous operation, where the machine performs all tasks by itself.
Compared to public roadways, the construction site is a comparatively controlled environment, with fewer conflicting vehicles or fast-changing environmental concerns. Nevertheless, reaching the goal of full AV autonomous operation in construction will still require overcoming many of the same technological challenges experienced by AV passenger vehicles.
For example, human eyes can see roughly four times farther in daylight than sensors. Rain, snow, dust or grime can inhibit camera, LIDAR or radar system sensitivity, thus reducing the AV software’s ability to assess and predict movements of other objects — site workers and other vehicles in particular. This has currently restricted fully automated AV functionality in construction to what are termed “short load cycles” — repetitive, loading-unloading interactions between the machine and an adjacent dump truck or conveyor belt.
In fact, a 2016 Swedish analysis detailing the problems facing AV in construction, titled Key Challenges in Automation of Earth-Moving Machines, concluded that even the complete automation of short loading cycles, “is not viable in the short to mid-term… The difficulty in automating the entire process can be attributed to the fact that it is impossible to accurately model the earth-moving process, especially the interaction between the tool and the environment.” The report said the central problem relates to the variety of media to be excavated, such as snow, soil, gravel, wood chips, fragmented rock, and mud.
One of the key attractions of full AV robotic operation is enhanced safety, specifically the possibility of removing site workers and operators from situations where they could be at risk. The mining industry has been successful in this regard, putting full autonomous operation into practice in specific highly controlled environments, allowing 24-hour operation in some cases. However, restricting or eliminating area access for construction workers is often not feasible.
The other key attraction of AV operations is the potential for productivity gains — a combination of loading-unloading cycle times, fill factors and fuel efficiency. However, while the Swedish report acknowledged that, “the loading step in the short loading cycle has greater potential for improvement with regard to the cycle time than navigation and dumping,” it also pointed out that shorter operational cycles increase the fuel consumption of equipment due to higher levels of acceleration and deceleration. That’s an important trade-off, considering fuel represents between 30 and 60 per cent of a machine’s operational costs.
Then there is the fill factor. While a built-in weigh scale system might be incorporated, the report said that more research is required because, for example, “bucket-rock interactions are much more complex than bucket-soil interactions.”
The Swedish report is, of course, three years old, and AV development in construction equipment has not stood still in the interim. Built Robotics based in San Francisco, for example, has developed self-driving earth-moving equipment that automates rough and finish grading, spreading, pushing and track walking, with safety features like wireless emergency stopping, geofencing confinement and sense-and-avoid technology.
Concurrently, the leading global manufacturers continue to release their own new automated developments, many focussed on semi-autonomous and tele-remote functions for equipment large and small. These are quickly being adopted by the construction industry, confirming a demand for increased automation.
John Bleasby is a Coldwater, Ont. based freelance writer. Send comments and Inside Innovation column ideas to firstname.lastname@example.org.