Development and Implementation of an AI Backed Human Detection Safety System for Pilings Rigs and Cranes

Contact with moving plant is one of the most significant health and safety hazards on modern construction sites. This risk can be exacerbated when employing specialist piling equipment with restricted operator visibility. To help reduce this risk, an AI backed human form detection and recognition camera system has been development and installed on piling rigs and cranes. The system recognises people or objects if they get within designated configurable exclusion zones and automatically suspends high risk operations of the machine, such as tracking and slewing. The system uses a combination of operator aids and audible exterior warnings dependant on the detected situation. This paper will present the authors experience of the evolution of rig/crane operator aid technology over the last 20 years, starting with simple reversing cameras to the first use of 360 degree camera systems, culminating in the latest AI backed human detection system currently undergoing development and trials by Cementation Skanska in the UK, developed in conjunction with Spillard Safety Systems. The paper will provide an overview of the new system and share some of the key challenges in its development and incorporation on to specialist piling plant. It will also discuss the human behavioural challenges that such systems have to overcome to ensure they are fully supported and adopted by the end user.