Artificial intelligence is more than just a buzzword. When applied to video telematics systems, AI is helping fleet administrators curtail risky driving like never before. AI-based systems deploy camera (or machine) vision, machine learning, edge computing, and other data-based technology to more quickly and accurately identify risky driving, analyze it, and deliver data to fleet management dashboards. Fleet administrators are using these video clips and more accurate driver scoring to improve and automate the coaching of drivers. This helps drivers, who can now curtail their risky behavior on the fly with real-time audible alerts that go beyond a series of beeps. At its most basic level, AI simulates human intelligence in machines – including dash camera systems – that are programmed to mimic the thought patterns (known as neural networks) and actions of humans.
When large data sets are fed into the algorithms that run the AI, the cameras become more capable at recognizing patterns on the road ahead, as well as inside the truck’s cab. By more quickly and accurately recognizing risky driving, AI systems provide video clips that provide much greater visibility into a fleet’s operation, allowing fleet administrators to more confidently correct problematic driving. AI-dash cameras represent the latest iteration in video telematics that began with manually activated cameras with limited onboard storage and evolved to the more robust video telematics solutions that are now the industry standard.
An AI-powered video system also provides greater automation for driver coaching workflows to lower the administrative burden and foster self-coaching by drivers.
Camera Vision: A Technology Boost Over Video Telematics
AI systems can help fleet managers:
• Review risky driving clips more quickly
• Provide more accurate video clips
• Identify distracted driving behaviors
• Improve driver performance
• Automate driver coaching with real-time alerts
• Lower risk
• Protect the company against false legal claims
AI-based camera systems represent an upgrade over non-AI systems that rely on G-force triggers from a device’s accelerometer to report speeding, hard cornering, and other fundamental risky driving events.
On a driver-facing lens, camera vision helps the device differentiate between a series of risky behaviors such as distraction, drowsiness, smoking, eating or drinking, and seat belt use. It also enables driver login and facial recognition.
On a road-facing lens, camera vision helps the device with a concept known as object detection – the camera becomes more adept at recognizing the difference between a stop sign, pedestrian, or other vehicle.
Machine Learning: The Power of Data to Solve Driving Problems
AI-based video systems rely heavily on data. In fact, the more data that comes into play, the more accurate the device. AI software continuously learns by processing data to improve the results of the events they recognize.
The advanced systems deliver more accurate triggers around risky driving events such as speeding, following distance, and distraction to reduce false-positive alerts. In the case of speeding, an AI camera will improve detection because it reads signs with posted limits to supplement map data.
Machine learning also improves each iteration of the device because more and more data improve pattern recognition. Machine learning helps fleet video systems learn patterns so they can provide alerts without human intervention.
Edge Devices: More Visibility into Fleet Operations
AI systems process driving data at the device level in the cab (“on the edge”) or apply AI in the cloud to analyze video footage. Whether the AI is applied in the cab or on the cloud, the analysis is done more quickly and accurately. On a non-AI system, the camera records footage and sends it wirelessly to a cloud server to be analyzed, coded, and made available at a later date. The edge camera captures much more video data, fleets gain a fuller picture of risky events. While some companies prioritize computing power on the device, others seek a balance between the edge and the cloud to keep a lid on data-transmission costs. A state-of-the-art AI-powered edge device needs high-end GPUs (graphics processing units) or DSPs (digital signal processors) on the camera’s chipset, which can boost the price of the hardware. AI In the cloud offers plenty of benefits because the analysis is being applied to video snippets that have been captured and uploaded. The snippets can be coded for review by the fleet administrator.
Automation: How Fleets Can Reduce Fleet Management Burden
One of the most significant benefits of AI-powered video systems relates to the driver coaching workflow. At their essence, video telematics systems capture risky driving behavior in the form of event videos – clips of speeding, hard braking, close following, and distraction. They transmit these clips as events to the fleet management platform, where an administrator can view them.
Fleet owners and managers can use the event snippets to reduce risky driving in their driver pool by coaching the riskiest drivers to reduce behaviors that may cause an accident. AI-based video systems help fleets automate their coaching workflow, by providing in-cab voice prompts that encourage the driver to alter behavior in real-time rather than waiting for a weekly or monthly coaching session with a supervisor.
The alerts usually go beyond the typical beeps of legacy telematics systems to include audible prompts to the driver around the risky behavior. These spoken alerts can nudge the driver to begin altering behavior on his own. To streamline driver coaching, fleet administrators can save time when they manage by exception – focusing on the lowest- and highest-performing drivers. Risk management among fleets entails identifying and scoring each driver and categorizing them into baskets of similar drivers. With this approach, high-performing drivers who represent the lowest risk to fleets can be signaled out for recognitions and awards. Fleet managers can also focus their attention on drivers who represent the highest risk to a fleet. It’s become a familiar refrain for fleet and safety managers that their lowest-performing drivers represent a persistent issue that threatens to boost a company’s accident rate and increase pressure from the commercial insurer.
AI can identify dangerous driving situations early enough for the driver to respond safely.
AI Helps Manage Fleet Data More Effectively
Fleet managers often struggle with managing too many streams of data from a multitude of sources, including vehicles, drivers, as well as back-office workflows around estimating, invoicing, and billing. When you consider the burden of complying with federal transportation regulations for driving hours and safety, the thought of adding more information from video systems could become too much.
AI helps streamline video review because it provides more accurate and clear insights on the manager dashboard. It reduces the amount of time fleet managers need to dig through video clips looking for specific events. AI-powered systems also provide event video in near-real-time or provide live streaming in some format. When video footage is available almost instantly following a collision, a fleet manager can speed up the claims management process by supplying needed clips to the company’s insurer.
Next Steps for AI-Powered Video
More and more video-based fleet management systems have integrated AI in the past several years, as the next iteration from video telematics systems.
Automating driver coaching workflow helps fleet, especially field service management businesses, do more with constrained resources. The next five years should see significantly improved commercial driving safety records with AI advancements.