“People Out of the Risk Zone”: A Unified System for Safe Industrial Robotization

Russia’s push to increase industrial robotization density directly intersects with a fundamental goal of occupational safety: removing people from exposure to workplace hazards. Statistics show that 96% of incidents stem from unsafe human behavior, while upstream facilities record between one and three thousand near misses annually, mostly due to failure to wear PPE. With manual oversight, detection depends on a specialist happening to be in the right place at the right time, and human patrol operators cannot continuously monitor tens of thousands of square meters of oilfield facilities. Aleksey Mosyagin, Head of HSE Service at ZARUBEZHNEFT-Dobycha Kharyaga LLC, examines the experience of building an automated safety system at an oilfield in the Nenets Autonomous Okrug.

Unified Architecture: Video Analytics, Inspection Robot, and Droneports

The novelty of this approach lies in combining fixed cameras, ground platforms, and unmanned aerial vehicles into a single AI-managed network with unified detectors and data analysis algorithms.

• Fixed video analytics and access control. Across the facilities, 163 cameras are deployed (covering 50% of the territory) to detect fire, smoke, PPE completeness and proper use (including hard hats worn backwards), and person-down events. When smoke or a fallen worker is detected, an alert is instantly sent to the Central Engineering and Technical Service (CITS) shift supervisor to dispatch emergency responders. Integration with the access control system (ACS) and facial recognition automates disciplinary workflows: if a worker lacks required gear at the entrance, the turnstile locks immediately, and if a violation is recorded on-site, the employee’s or contractor’s badge is blocked for the next shift pending an investigation. Additionally, checkpoints feature thermal screening with secondary manual verification by security guards to prevent ill personnel from starting their rotation.
• Mobile inspection robot. A platform equipped with LiDAR, 360-degree cameras, a laser methane detector, and a diffusion hydrogen sulfide sensor patrols the central oil gathering facility. The robot covers an additional 10% of the area outside the view of fixed cameras, monitoring gas levels and PPE compliance. A dedicated scenario involves reconnaissance of a 100-meter cordon zone following UAV debris impacts, where human entry is prohibited until EOD specialists arrive. The robot enables remote situational assessment so firefighting can begin safely.
• Autonomous droneports. Monthly manned pipeline overflights and tracked vehicle patrols across marshy tundra pose risks to crews and cause delays: within a month, a micro-leak can escalate into a major spill. Deploying two automated droneports (scheduled by February 2027) will provide continuous thermal imaging monitoring of linear assets for early failure detection.

Year-Long Arctic Robot Trials: Breakdown of Technical Challenges

The presentation examines in detail nine categories of issues identified during a year of pilot operation of the inspection robot—factors critical to include when drafting vendor technical specifications:

• Navigation and communication failures. Wi-Fi and GPS signal loss caused the robot to stop in the middle of roadways for up to 30 minutes, obstructing heavy equipment traffic, while positioning errors risked sending it off-route. Additionally, a mechanical rightward chassis drift forced constant steering corrections and accelerated battery drain.
• Computer vision errors. On clean snow in sunny weather, glare blinded the optics, while in darkness the system detected phantom obstacles. Baseline algorithms failed to recognize barrier tape and temporary power tool cables mounted on tripods, creating a risk of driving into hazardous zones or entangling cables.
• Cold-weather failures. At temperatures below −30 °C, the unit failed to launch from its outdoor charging station (requiring thawing in a heated bay), the methane sensor showed measurement drift, and straight-line speed dropped alongside accelerated mechanical wear. Actual snow mobility reached only 10 cm instead of the rated 20 cm.

Five Stages of Safe Robot Integration and Project Economics

Because robots themselves introduce technological hazards for personnel, the company analyzed safeguarding solutions used by oil and gas peers (magnetic interlocks, laser area scanners, RFID tags, and the requirements of the new industry standard GOST R 60.1.8-2026) and built a five-stage implementation framework. It includes: analyzing process incident rates to select areas offering maximum risk reduction; evaluating payload capacity, ingress protection, software compatibility, and docking accuracy; designing a safe work environment with physical and warning barriers; developing internal regulations, occupational safety instructions, and hands-on operator training under Government Decree No. 2464; and conducting pilot commercial operation.

The inspection robot was deployed without headcount reductions, freeing up 2.37 FTEs from routine patrols for analytical tasks. Payroll-equivalent savings reached 7.7 million rubles per year, with a payback period of 2 to 2.5 years and zero robot-related incidents. Under the roadmap through 2030, the company plans to deploy 12 more industrial robotic systems, including robots for tank and pipe diagnostics and warehouse operations.

What You Will Learn from This Presentation

• How can fixed video analytics, access control systems, ground robots, and drones be integrated into a unified automated safety monitoring system?
• What navigation, optical, and mechanical failures does an inspection robot face in temperatures down to −32 °C, and what must be written into vendor specifications upfront?
• How do you legally and technically set up automated badge blocking for employees and contractors who fail to wear PPE?
• What engineering barriers and procedures protect personnel from injury caused by industrial robots themselves?
• How do you calculate the economic return and payback period of robotization through freed-up specialist time?