An Advanced Fleet Management System For Trucking Companies

An advanced fleet
management
system for
trucking
companies

Get in touch

Client

Intel

Industries

Automotive

Services

Solution engineering

Technologies

Cloud

Introduction

As a reputable software development vendor with significant expertise in the automotive industry, Avenga has consistently strived to push the boundaries in the areas of transportation and fleet management.

With this objective in mind, we conduct thorough research and development activities in the automotive field, focusing on designing and engineering cutting-edge solutions that enhance driving safety and provide trucking companies with a competitive advantage. We have also collaborated with many industry-leading software companies, including Intel, on numerous automotive projects in the automotive and other industries.

Notably, one of our fleet management initiatives has generated interest from local municipalities in Cordoba, Argentina, where one of our delivery centers is based. The work we have undertaken demonstrates considerable potential in terms of elevating driver monitoring, improving driver safety, and increasing fleet management efficiency.

Challenge

With cities worldwide undergoing rapid development, trucking companies are under mounting pressure to enhance fleet management, including tracking and coordinating driver and fleet behaviors.

To address modern transportation needs, they need sophisticated fleet management tools, which could help streamline processes, prevent dangerous situations and, ultimately, increase their competitive advantage. In light of this, Avenga has decided to engineer an ML-based, non-ADAS tool tailored to these industry-specific requirements.

Solution

Innovative POC

Engineered a proof of concept (POC) for a sophisticated system that proactively notifies the driver of any potential emergencies or dangerous situations and tracks their behavior. The system's innovative architecture is designed to leverage the combined capabilities of cloud computing and edge processing, providing a robust and scalable platform for managing and analyzing large volumes of data in real-time.

Intel’s OpenVino toolkit utilization

Through the utilization of Intel’s OpenVino advanced toolkit, we were able to implement highly precise detection capabilities within the platform. The solution’s design incorporates both facial recognition and gesture detection modules, which enable real-time identification of varying levels of driver activity and rapid detection of potential signs of drowsiness or distraction.

Robust fatigue detection functionality

Overall, we developed an advanced system for detecting driver fatigue that monitors driver behavior and assesses the risk of fatigue-related incidents. The system leverages facial landmark detection to analyze the driver's eyes and mouth, identifying signs of fatigue such as prolonged blinks and yawns. Additionally, we added a gaze attention monitoring mechanism that determines whether the driver's face is centered or not centered and also helps reduce the risk of accidents.

Event recording

If the system detects that certain behaviors have exceeded their defined limits, it provides the user with two options. The first option is a VU meter that records the number and persistence of these behaviors within specific time frames. The second option is a tool that records the events and allows trucking organizations to store them in the cloud for future driver coaching.

Powerful tech stack

We utilized Amazon Web Services to ensure the required computations could be conducted at the edge, inside the vehicle, thereby eliminating the need for cloud computing and large servers. Furthermore, we skillfully implemented the combination of AWS Elasticsearch, Logstash, and Kibana (ELK) to facilitate rapid data processing and enable the creation of dynamic dashboards, which enable efficient driver tracking, behavior assessment, and support.

Near-misses system

Additionally, we engineered a near misses detection system that relies on AI to accurately detect vehicles and pedestrians at intersections in cities. The tool is designed to work with smart cameras and can be used for indoor and outdoor monitoring. It can integrate with external cameras and merge data from multiple sources, which enables it to track objects within predefined areas of interest with high precision. It also can process data at the edge and stream it to the cloud for structuring and analysis.

Results

We delivered a proof-of-concept (PoC) for an end-to-end fleet management solution that can optimize trucking companies’ processes, improve driver safety, and enhance their competitive edge.
This PoC has already gained the attention of the municipal authorities in Cordoba, Argentina, where our delivery center is located, and demonstrated significant potential for real-world applications.
We also engineered a powerful near misses detection system that employs machine learning algorithms to identify potentially hazardous situations at city intersections.
Finally, we achieved the integration of driver management and the near detection systems into Intel’s IoT DevCloud platform.

Technologies

AWS, Elasticsearch, Logstash, Kibana.

Frequently Asked Questions

What are the key advantages of Avenga's fleet management system?

The proof of concept (POC) we have developed exhibits substantial potential for a wide range of real-world use cases. Specifically, it can assist truck drivers and help trucking companies address the changing transportation demands.

How was Intel’s OpenVino toolkit used in this project?

Utilizing Intel's OpenVino toolkit in this project was advantageous in various aspects. In particular, It helped expedite the development process and maximize the solution’s performance.

What technologies did you utilize for the platform?

Our tech stack consisted of a variety of carefully selected tools. Notably, we leveraged Amazon Web Services, specifically AWS Elasticsearch, Logstash, and Kibana (ELK).

What is the functionality of the near-misses system?

Based on the OpenVINO toolkit, the near-misses system aims to enable the deployment of AI models for efficient and accurate detection of vehicles and pedestrians at intersections within cities. It can help city planners prevent potential accidents.

Intel: An advanced end-to-end fleet management system