After three of his brothers were seriously injured by construction site vehicles at work, Dr. Nathan Kirchner decided it was time for the industry to change.
Together with Dr. Kieran MacKenzie he founded tech start-up Presien – a Laing O’Rourke spinout, aimed at preventing vehicular accidents in heavy industries.
Kirchner’s personal experience was not the only motivation for forming the company. More than 182 Australians were fatally injured at work last year – a disproportionate number from industries, such as construction, mining and logistics. This is despite these sectors being heavily mandated in terms of safety protocol and equipment.
While traditional safety tech in vehicles, like rear-view cameras, can cut crashes by up to 78 percent, this figure is believed to be far lower in worksite vehicles, for one simple reason: the vehicle operator can’t give their full attention to the camera footage.
“People seem to forget that a person operating a vehicle in a worksite isn’t just driving it, they are simultaneously performing a job, like digging a hole or removing scraps,” said Dr. MacKenzie.
“It is all very well having state of the art camera systems in place, providing a 360 degree view of your site. But camera footage is only effective insofar as the person is looking at it.”
Even momentary inattention can be lethal when operating heavy machinery. A study by the USA National Highway Traffic Safety Administration (NHTSA) found that around 80 percent of workplace crashes and 65 percent of near-crashes involved driver inattention, within three seconds before the incident.
Recognising the inadequacy of cameras in preventing these incidents, MacKenzie and Kirchner’s Presien tool can be programmed to detect objects of choice in a worksite and – using artificial intelligence – provide an alert to the operator when the object is in near range.
The ‘choice’ element is one of the key sell-points, they say, with superfluous alerts (to non-hazardous objects) likely to desensitise operators to the warnings, rendering them ineffective over time.
“You can program the AI tool to only recognise people if you wish, rather than objects, like piles of debris. That way, when your alert goes off, you know you have to pay attention to it, rather than dismissing it as a false alarm. It’s a complex design, but this complexity is hidden from the user, with a simple interface,” said MacKenzie.
An additional sell-point is the tool’s standalone capabilities.
“The tool doesn’t need internet connectivity to recognise objects and generate alerts,” said MacKenzie.
“This is really important for two keys reasons. Firstly, it means that the system can be used without reliable internet, which is common in heavy industry operations. Secondly, it stops latency issues, with the round-trip from the device to server processing too tardy for it to be effective as a hazard alert.”
The tool does, however, have inbuilt 3/4G capabilities to collate data for monitoring and investigation purposes.
“The goal is not only to stop accidents, but also to automate reporting. The object detection data, including video, can be automatically sent to the cloud, allowing users to better understand how risky their set-up is. Additionally, this data can be used for investigations in the event of any near-misses or accidents,” said MacKenzie.
The first version of Presien’s AI tool has already made a large footprint in heavy industries throughout Australia. And, more recently, the firm has secured $4million of venture capital funding from Main Sequence Ventures to further develop the technology.
Dr. Kieran MacKenzie – who has a PhD in Engineering and a decades’ experience in construction innovation – will share details of his workplace safety learnings at the RISSB’s Rail Safety Conference – held as a virtual event on 27-28 October 2020.
Learn more and register.