Improving the safety of construction machinery based on automated recognition of obstacles and hazardous zones

Abstract

Occupational injury indicators returned to growth in 2024: the number of registered accidents reached 33,774 at an incidence rate of 0.69, while the impact of moving objects and parts together with transport incidents accounted for more than a third of accidents with severe consequences. Current visibility requirements for earth-moving machinery regulate the local dimensions of masking areas and leave the total share of space uncontrolled by the operator unregulated, so that machines formally complying with the standard differ in actual coverage by almost one third. The aim of the work is to establish a quantitative relationship between the parameters of an automated obstacle recognition system and the residual probability of contact between a worker and a moving machine. Geometric modelling of masking sectors within a visibility test circle of 12 m radius, additive decomposition of the guaranteed actuation distance and simulation modelling over 20,000 realisations were performed for a tracked excavator, a wheel loader and a motor grader in four equipment configurations. Base configurations leave between 23,17 and 30,07% of the test circle area uncontrolled; an around-view system reduces the figure to 4,28-4,83%, and a combined system with a radar channel to 1,60-1,88%. The residual share of contacts decreases from 0,1840 to 0,0871 and 0,0820 respectively, which corresponds to a risk reduction of 52,7 and 55,4%. Within the required actuation distance at 9 km/h the operator reaction accounts for 51,6% and the latency of the computing path for 7,2%. The obtained distribution shows that the efficiency limit of the system is set by coverage geometry and by the reaction time of the deciding element, whereas detector speed affects the result about three times less. Once the uncontrolled share is reduced to approximately 5%, further expansion of sensor equipment yields diminishing returns, and the residual risk concentrates on the class of workers in a bent posture and on the upper range of operating speeds.