Factory-Fit Human Detection Systems for Crawler Excavators

Komatsu integrates Safety Shield Global technology to provide real-time human detection on crawler excavators.  www.komatsu.com Komatsu has launched a factory-fit Human Detection System as an optional integration for selected UK-built crawler excavators. Developed in collaboration with Safety Shield Global, the system is designed to identify personnel in the immediate vicinity of heavy machinery, providing operators with real-time alerts to mitigate collision risks on construction sites. Hardware Integration and Alert Mechanisms The human detection software integrates directly with the existing KomVision camera network and the excavator's onboard controllers. The architecture utilizes a dedicated in-cab display and an audible alarm system to notify the operator immediately when an individual breaches a predefined risk zone surrounding the machine. The vision algorithms have been calibrated and tested to reliably identify human forms across a wide variety of high-visibility and protective clothing colors typically encountered in industrial and construction environments. Factory Calibration and Platform Availability By offering the system as a factory-fit option, the integration process avoids the complexities and potential electrical incompatibilities associated with aftermarket retrofitting. The sensors and processing modules undergo factory calibration and validation alongside the machine's primary control systems. The Human Detection System is currently available as an option on several Komatsu models, including the PC170-11, PC210-11, PC240-11, PC290-11, PC360-11, the hybrid HB365-3, and the heavy-duty PC490-11. Additional Context: This section details technical specifications not included in the original announcement In heavy machinery automation, integrating human detection into an existing 360-degree surround-view system (like KomVision) requires advanced edge computing capabilities. Standard surround-view systems merely stitch together multiple camera feeds to present a passive bird's-eye view to the operator. To achieve active human detection, the system must employ machine vision algorithms, typically Convolutional Neural Networks (CNNs), running on a dedicated Video Processing Unit (VPU). These algorithms are trained to differentiate human kinematic profiles from static job site objects, such as traffic cones, concrete barriers, or soil mounds. When a human form is detected within the predefined exclusion zone, the edge computer sends a high-priority interrupt signal over the machine's Controller Area Network (CAN bus) to trigger the in-cab alarms with ultra-low latency, ensuring the operator has sufficient reaction time to halt the machine's swing or travel path. Edited by Lekshman Ramdas, Induportals editor – adapted by AI. www.komatsu.com Powered by Induportals Media Publishing