Over the years, the transportation industry has seen significant innovation, especially with the rise of vision-based driver safety platforms. Early vision-based driver safety platforms combined collision detection, video recording, and internet connectivity, to capture records of road accidents. These records were then reviewed by safety managers for accuracy and shared with drivers as part of a coaching program.
The next generation of vision-based driver safety platforms, starting with the introduction of Driveri in 2016, replaced manual review of possible collision video with continuous analysis of driving scenes using Artificial Intelligence embedded directly into the edge computing device. With an “AI” dashcam, a safety manager would now have access to every driving minute of a driver’s day, instead of the few minutes of driving minutes per month offered by the earlier systems.
When you start to analyze more complex data for millions of transport cycles, people, accidents, and road miles you move into the realm of Big Data. Based on continuous analysis of roughly 10 million miles of new driving data every week, the Driveri system has employed Machine Learning techniques to create and improve predictive analytics that boost driver performance in real-time and prevent crashes.
The Rise of Big Data in Transportation
The use of big data in transportation is rapidly growing. Driveri algorithms have been created, trained and tested on the big data streams its devices collect to make driver systems smarter.
Netradyne has been able to automate the process of big data collection, analysis, and application of the insights to improve driver performance so that the system continuously trains itself on the best way to optimize driver performance using past data.
Machine Learning and Artificial Intelligence
Machine learning is a subset of artificial intelligence (AI) that gives computer systems the ability to learn from data without being programmed continuously.
For example, you may want to develop a program that moves your car when the light is green and stops it when it is red. A machine learning engineer can create a short program that makes this task possible: (IF LIGHT=GREEN THEN GO, ELSE STOP).
This seems like a fairly simple problem to solve — until you consider all the variables that come into play. For example, what should happen when another car runs a stoplight? Now you have to make your program a little bit longer: (IF LIGHT=GREEN AND (NO CAR RUNNING THE OTHER RED LIGHT), THEN GO).
But what if your car is stopped at a traffic light that isn’t functioning, and other cars in front of you are running the red light on purpose? How will your vehicle calculate the right move at a crosswalk?
Instead of continuously adding new lines of code to your program to account for every variable, machine learning trains your program to recognize these variables and act on them, all by itself.
Machine learning begins with the collection of data from different sources. For example, Driveri collects data relevant to driver performance and changing road conditions. Through machine learning, its algorithms focus on which actions recorded in the data tend to positively influence driver performance, management, and road accident prevention in the face of changing road conditions.
The goal is to allow the system to learn continuously and at a scale that would be impossible to achieve if held-back by manual review cycles.
Artificially intelligent devices are essential for tasks that involve processing huge amounts of data, gaining insights from it and making the right decisions. This is something that humans do slowly and with only small bits of data at a time — giving technology a distinct advantage.
In the case of Driveri, we ensure that data is collected for every minute of every road journey. Our algorithms process this data and offer insights in real-time, instead of having a fleet manager review hours of video. The speed of data processing is crucial to gaining access to eight hours of training data per driver per day, rather than the few minutes of driving data per driver per month that legacy systems collect.
A typical commercial driver logs 100,000 miles of driving every year. Over a thirty-year career, an active commercial driver will have driven 3 million miles. Currently, Driveri systems collect that much data every few days. Because the AI is on the device, all of these miles are being analyzed and used to make the system smarter. At this scale, the infrastructure that Netradyne has built enables several lifetimes of human driving experience to impact every iteration of its increasingly intelligent AI processing system.
In the Context of Transportation
Driveri uses machine learning to train its platform to recognize distracted driving behaviors and risky situations, as well as positive and proactive driving events that the system seeks to reinforce. The system also analyzes road conditions, following distances, and other environmental factors to help your driver make the safest choice. Its algorithms work behind the scenes to drive reward systems, error recognition, and driver coaching.
In transport, AI solutions are gradually becoming an industry standard due to the rise of platforms like Driveri. These platforms can learn about their environments and intelligently apply what they learn. In an industry weighed down by accidents, insurance costs, and labor shortage, artificial intelligence makes it possible to harvest the piling data and propose solutions to these problems by reducing the frequency of accidents. Fewer crashes, in turn, decrease insurance premiums, all while engaging with drivers in a way that promotes long-term job satisfaction.
Driveri coaches drivers on their riskiest driving habits, showing them the best way to mitigate risks without a supervisor breathing down their necks. Large fleet managers can also be sure that the whole process of detecting poor behavior in drivers can be done in the background while they handle other pressing tasks.
Driveri Technology and Artificial Intelligence
Artificial intelligence forms the basis of the Driveri platform. Mostly intended for advanced data analytics, the technology trickles down into other features such as communication, and a rewards system.
The platform operates in three major steps: collecting data, interpreting it, and acting on it through three different sections. These sections are its vision-based data collection fleet camera fleet camera system comprised of cameras and sensors, its data analytics system using artificial intelligence, and rewards systems DriverStar and GreenZone.
Driveri Data Analytics System
The Driveri data analytics system collects insights concerning road conditions, traffic hazards, and driver performance, including distracted driving behaviors. After analyzing the data, the system coaches drivers through challenging situations and risky behaviors through real-time notifications. This can save fleets millions of dollars in damages by averting potential accidents.
Vision-Based System
Driveri’s vision-based system is primarily used for data collection while on the road. Internal and external cameras capture driver behavior, capture accidents as well as road altercations, and capture road data. The system consists of a group of side, front-facing, and interior QUAD HD cameras for high-quality video recording and real-time event playback. Managers can access up to 50 hours of video playback when needed.
Driveri combines AI and machine learning to create and operate its artificially intelligent driver management platform which collects and analyzes data in real-time. With the help of technology, drivers can mitigate risks, while fleets can maintain safety and monitor driver performance in real-time. Contact us to learn more about our advanced Driveri technology.
On-Demand Webinar
