Managing risks and protecting workers by scaling AI in near real time
- $1.6M Savings in cargo theft
- 10M Queries per month
- 271K+ Tons transported per month
G2L Logistica (G2L) has become one of the largest logistics operators in Brazil by leveraging cutting-edge technology. G2L began as a pilot project at Gerdau, Brazil’s largest steel producer, and has grown over 460% in five years, transporting much of the nation’s steel, chemicals, construction materials, food, and consumer goods.
For G2L, worker safety takes precedence over all other objectives and priorities. No delivery or business result should compromise the health or safety of G2L’s workers, the environment, or the quality of services provided. As part of this commitment, G2L implemented AI, analytics, and data solutions to improve worker safety, mitigate risk, and deliver value to customers.
Using data and IoT to reinforce worker safety
This commitment comes with challenges. Brazil’s geographical size and infrastructure presents hurdles for logistics operations. A mix of aging and new infrastructure creates traffic congestion, necessitates long-haul transport, and sometimes requires driving on unpaved roads and in remote locations. All of these factors can present safety risks.
“Brazil is a continental country,” says Paul Seiko Higa, data engineer at G2L. “We have more than one million roads, and we must transport heavy loads across the country, where the infrastructure isn’t the best.”
Despite this, G2L’s 1,800 drivers transport over 271,000 tons per month with over 300 vehicles in its fleet, supporting 4,000 Brazilian cities. Woven into G2L’s culture is a belief that all business problems can be solved with analytics and data. G2L has digitally enabled services, such as a control tower and end-to-end cargo tracking visibility, powered by internet of things (IoT) geolocation data on its fleet, cargo, and drivers.
“We use IoT geolocation data from our trucks to track the routes,” says Higa. “We package this along with information from all our departments (financial, human resources, operational) in Teradata VantageCloud Lake on Microsoft Azure. That allows all sectors of the company to understand how we’re performing and make decisions better and faster.”
Some of G2L’s most pressing use cases for AI fueled by unstructured IoT data include:
Engineering
Worker safety
Improving safety standards for drivers to reduce roadside accidents
Health_And_Safety
Workplace accidents
Reducing workplace injuries during cargo loading and unloading
Money_Bag
Cargo theft
Reducing financial loss and driver exposure to unsafe conditions in remote locations
AI and VantageCloud Lake protect employees above all
For G2L, operational safety is nonnegotiable—so AI and data present crucial opportunities to protect G2L team members and the communities G2L serves.
“Cargo theft is a problem in Brazil because the roads are very long,” says Higa. “Cities are far from one another, and there are places with few people. This can lead to risks like cargo robberies and driver assaults and kidnappings.”
By deploying geolocation IoT devices in its vehicles and geofencing routes, G2L’s Central Command Center monitors vehicles in near real time. Central Command sees when vehicles make unscheduled stops or deviate from planned and approved routes.
“Python scripts connect to an application programming interface that receives the IoT geolocation data, which streams into VantageCloud Lake,” says Higa. “Central Command can see each truck’s location and destination—and monitor to prevent risks and theft attempts,” explains Higa.
Before implementing VantageCloud Lake on Azure, G2L faced challenges with receiving timely and actionable data. “We could only acquire data in 30 minutes,” says Higa.
Using VantageCloud Lake to stream near-real-time data and monitor driver routes has saved lives and mitigated risks. “With VantageCloud Lake data in near real time, we didn’t have any robberies this year, and we’re saving $1.6 million compared to last year,” says Higa. “Best of all, none of our drivers had any problems with assaults, kidnappings, or robberies. Keeping them safe is the best part.”
Maximizing data innovation with a lake-centric modernization approach
G2L also sought to manage its dispersed and decentralized data, which led to poor visibility and inadequate responses to business challenges. Inconsistent data with little governance often resulted in poor performance and high costs. Implementing AI use cases was challenging due to limited availability of data science tools.
“We had infrastructure that grew too fast and wasn’t scalable,” says Higa. “We had broken jobs and inconsistent data with no single source of truth. Since implementing VantageCloud Lake, we’ve grown our queries from 70,000 to 10 million per month. This is because we can now acquire data in two minutes, which allows us to track our trucks with the hot, more precise data. It’s much more stable, and we can scale up and down and control costs."
Additionally, G2L can now perform AI and advanced analytics with ClearScape Analytics™ for structured and unstructured data. This supports the use of image and video data to advance driver safety and reduce workplace accidents. Drivers face high fatigue rates on long hauls, and fatigue can lead to accidents.
To overcome this, G2L is now implementing IoT cabin monitoring devices and using video analytics models for driver behavior detection. If G2L can identify driver fatigue or other high-risk behavior patterns, it can alert drivers in real time and deploy safety measures to reduce accidents.
“With VantageCloud Lake, we can store and process unstructured data, such as in-vehicle driver footage, drone footage, and images,” says Higa. “We can then use ClearScape Analytics to train our models.”
G2L plans to expand its use of AI-powered video analytics to its distribution centers and branches. “Cargo loading and unloading presents safety risks,” Higa says. “By using visual AI to analyze videos and pictures of cargo fastening, we can monitor and assure we’re securing cargo correctly.”
Using drone footage and pattern detection AI models, G2L aims to reduce incorrect loading techniques, improperly fastened cargo, and other personnel behavior that can lead to accidents and injuries.
“Our focus is to prevent problems and create safer working conditions,” Higa says. “We can do this using machine learning and AI for driver safety and to prevent operational accidents. Ultimately, we can provide the business with information to make decisions that improve the lives of our workers.”
Connectons-nous
Découvrez comment Teradata VantageCloud peut vous aider à accélérer les résultats commerciaux et à fournir l’agilité dont vous avez besoin.
Nos représentants commerciaux sont là pour vous aider.