Vehicle Monitoring System

Driving a car is a process that quickly becomes routine for a person. Because of this, drivers often stop perceiving driving as a high-risk job. To change attitudes towards road safety, companies traditionally use two tools: defensive driving as a method of persuasion and on-board monitoring systems as a barrier. In his presentation, Evgeny Miroshnichenko, Head of the Industrial Safety Automation Program at Gazprom Neft PJSC, examines the evolution of vehicle control systems from basic telematics to intelligent complexes with video analytics.

Symbiosis of a Digital System and a Live Dispatcher

A key insight during the implementation of the vehicle monitoring system was the realization that in the early stages, a digital tool needs a human. Strict automated control is often perceived by drivers as a punitive mechanism, which causes rejection. Introducing the role of a dispatcher who contacts the driver to clarify the situation and provide assistance radically changes this perception.

The speaker demonstrates this using a pilot project covering 250 heavy-duty vehicles transporting hazardous materials: adding the human factor to the dispatching process reduced the total driving time with speeding violations from 23 to 6 hours per week in just five weeks. The dispatcher works with a smart event feed, where the system automatically prioritizes incidents based on weather conditions, time of day, and type of violation, allowing the operator to focus on the highest risks.

Matrix of Measures and Evaluation of Managers' Actions

Recording a violation is only the first step. How management reacts to it is much more important. The presentation details the escalation mechanism and working with the matrix of measures. If the dispatcher cannot contact the driver, the signal is transmitted to the line manager, and in critical cases, to the general director of the enterprise.

The system automatically generates recommendations for applying disciplinary measures depending on the severity and recurrence of violations. At the same time, the manager receives a link to the event card and can either agree with the system's recommendation or choose their own measure. This approach has a dual effect: the company gets not only a tool for managing driver behavior but also an objective metric of the managers' own engagement, tracking whether they ignore incidents or actively work with subordinates.

Video Analytics and On-board Neural Networks

The next stage in the development of telematics is the transition to predictive analytics on board the vehicle. The speaker describes prototypes of compact computing modules that combine the functions of a dashcam and a neural network. This technology allows recording not only the vehicle's movement parameters but also the driver's dangerous actions: unfastened seat belt, smoking, phone use, signs of fatigue, and falling asleep.

To provide feedback, on-board computers are integrated with speech modules. Instead of incomprehensible sound signals, the driver receives clear voice instructions and warnings — for example, about entering a speed limit zone, the need to turn on headlights, or approaching a danger zone. This turns the system from a control tool into a full-fledged digital assistant.

What you will learn from this webinar:

• How does the integration of a live dispatcher into the monitoring system help to significantly reduce the number of speeding violations?
• By what algorithm does the system prioritize dangerous events for processing by the operator?
• How does an automated matrix of measures allow evaluating the effectiveness of line managers' work with drivers?
• What parameters of driver behavior can an on-board neural network with video analytics monitor?
• How does the use of speech modules change the driver's interaction with the on-board monitoring system?