Image: Mike2focus/Dreamstime.com

News

Construction site complexity will slow adoption of autonomous vehicles

14 July 2019 | By Stephen Cousins

Construction will lag behind other industries in the uptake of self-driving machines due to the inherent complexity and rugged conditions of job sites, latest research reveals.

The report by technology consultancy Arthur D Little, based on a survey of more than 30 industry and technology experts, names building as the “most difficult environment for autonomous technology”, due to the complex and dynamic conditions where off-highway vehicles are typically used for a limited amount of time at changing locations.

The vast majority of machines on construction sites will require human operators to assist with most driving and work functions (so-called “Level 2” automation) for the next 10 to 15 years, says the report, apart from in certain niche applications, such as haulage trucks.

The sector will lag behind mines, warehouses and other locations, where hundreds of vehicles already run driverless with Level 3 automation and are expected to advance in capability in the near future.

Alexander Krug, partner at Arthur D Little, commented: “The autonomous ‘pioneer’ applications of today and the close future all have similar characteristics, many of which are not fulfilled by the majority of construction machinery. They have a high utilisation duty cycle, but construction machines are often multi-purpose and have very low utilisation. They operate in a closed-off work environment, but building sites are complex and dynamic and machines that are only used for a limited time at changing locations.” 

He continued: “Many processes on a construction site require the exact interplay of several machines, which would require automation of the entire process.”

The prospects for driverless vehicles on civil road and underground engineering projects are more hopeful, says the report, due to more suitable use cases and site environments.

“Dump trucks and carriers will probably be automated first, then roller compactors for road building, followed by high power loaders with high utilisation, which meet the typical

characteristics of autonomous machines,” says Krug. “Other use cases, like excavators, are too complex and will either not become autonomous soon or their autonomy will be limited to a single and repetitive work function.”

Half of respondents quizzed by Arthur D Little said their company had already invested in R&D, demonstrators or first projects for autonomous technology, and 30 percent said they were already offering semi-autonomous products.

Key players involved in the development of autonomous off-highway vehicles for construction include OEMs like Volvo, Komatsu and Caterpillar and specialised robotic companies doing retrofits such as ASI and Built Robotics.

Autonomous machines offer numerous benefits such as 24/7 operation, higher productivity, reliability, reduced labor requirements and improved safety, by eliminating human error and taking people away from difficult and dangerous work environments.

However, they are currently too expensive for the majority of use cases in most industries, said the report, mainly due to the cost of high-performance components and systems, including sensors, software and high-performance back end.

Automation Levels explained

The degree of automation of driving and work functions in an off-highway vehicle is typically split into four distinct categories. 

Level 1 - No automation, 100% manual operation.

Level 2 -  Automation and guidance for single work functions, eg automatic steering. Still requires a human operator to assist with most functions.

Level 3 - Semi-autonomous. Automation of complete work processes and a combination of multiple work functions, eg short cycle loading on a construction site. These machines exploit 3D-perception sensors such as LiDAR, radar, and stereo cameras, global navigation satellite system and specific sensors to monitor custom work functions.

Level 4 - Fully autonomous with complete automation of the system. Able to perform a combination of multiple work processes thanks to greater sensor integration and software enhancements.

Image: Mike2focus/Dreamstime.com