ABOUT

AI & Video based telematics solutions address critical challenges in various industries by combining video recording technology with telematics data and automating it with the use of AI. These systems are essential for enhancing safety and minimizing risks associated with fleet management. By monitoring driver behaviour through video analytics, organizations can identify and rectify unsafe practices, leading to a reduction in accidents and improved road safety. The recorded footage serves as valuable evidence for incident reconstruction, aiding in insurance claims and mitigating liability. Additionally, video telematics contribute to operational efficiency by optimizing routes, promoting fuel savings, and facilitating proactive vehicle maintenance. The real-time monitoring and alert features offer instant notifications for events like rash or negligent driving, enabling prompt intervention to address issues and prevent potential risks.

Furthermore, video telematics solutions play a pivotal role in compliance management, ensuring adherence to industry regulations and standards. The technology also supports driver training initiatives, allowing organizations to provide targeted coaching based on performance data. This not only enhances the skills of drivers but also contributes to long-term safety improvements. Overall, video telematics solutions provide a comprehensive and integrated approach to fleet management, offering benefits such as insurance cost reduction, enhanced security, and streamlined operations, making them indispensable tools for organizations seeking to optimize their transportation and logistics processes.

INDIA

India, with its vast geography, terrain, and an expanding road infrastructure, has a need for solutions that help manage and address challenges that affect businesses in the transportation, logistics, construction and other similar services.

GPS based tracking devices, now a regulatory mandate for commercial vehicles, are widely available and used. The information captured by these telematics devices provide effective data points that help organizations identify driving patterns, track vehicles and gather vehicle specific data and metrics.

With market maturity, early adopters are keenly exploring solutions that are more sophisticated and provide tangible information that not only provides insights to driver and vehicle metrics but can also address and change driver behaviour in a positive way.
There are about 5 lakhs road accidents every year in India and 78% of those accidents are caused due to driver’s fault. Top causes of these accidents include and are not limited to the following:

• Over-Speeding
• Driving under influence
• Driving discipline – Wrong side, overtaking, lane switching
• Jumping red light
• Use of mobile phone
• Failure to wear a seatbelt

Source: Ministry of Road Transport and Highways.https://morth.nic.in/sites/default/files/RA_2021_Compressed.pdf

One aspect of the AI and Video Telematics solutions aims to touch on the above accident reasons by warning drivers, capturing event videos for driver training, and use recorded video as reference for managing operations, verifying incidents, evidence for insurance claims, etc.

DHL ENGAGEMENT

DHL India’s requirement primarily stemmed with the focus on Driver Safety and extension of company rules and policies to partners. As an organization, DHL has a company-wide policy on ensuring safety & security of not only its employees but also contractors and partners engaged with the company.

The decision for choosing the right partner for the telematics solution was based on several reasons and evaluation criteria. The key ones that differentiated ‘ZigFleet – AI & Video Telematics solution’, from the rest included

1. Compliances to different scenarios outlined and flexibility to add newer scenarios in the future
2. Structure the AI model to Indian road infrastructure and conditions.
3. Planned roadmap of services such as Digital Twin, Driver Coaching, etc.

CHALLENGES

Introduction of any new product that directly impacts the driver behaviour, driving pattern and influences change, has its challenges.

During implementation of ZigFleet, there were challenges that came in at different times, and some continue to be today too. These sets are applicable industry wide and not unique to this engagement alone.

• Hardware tampering: With the cameras in the vehicle cabin, the drivers felt intrusive of personal space. That leads to frequent issues of camera being turned away, removed, damaged or missing. User acceptance is crucial for the successful long-term implementation of these solutions.
• Road & Driving conditions: The road and driving conditions in India is not systematic and organized. Driver cabin conditions, low-light, jaywalking, wrong way driving, missing lane markings & more, all created a challenge for event detection and alerts for the driver. Furthermore, minor tweaks in road layout and traffic signals from city to city led to increase in false positive events for inter-state trips.
• Hardware limitation: After-market solutions have a limitation on the capability of the hardware. E.g. extreme environmental conditions, such as adverse weather or low-light situations. There needs to be a balance between cost vs benefit. Expectations like the examples listed above, requires a hardware with high specs, which directly translates to cost negating the value that the product offers.

The Solution

In line with the challenges faced, specific actions were taken to overcome them –

1. To address hardware tampering, two key approaches were taken (1) AI and sensor data was used to detect changes to camera angles, loss of power & connectivity to hardware (2) Proactively educating the drivers that the cameras were installed to enhance their safety and not to penalize, resulting in reduction of related issues.
2. A customizable AI Model was used to cater to local traffic conditions and geographical requirements. Periodic updates were planned to tune the AI model to accommodate new unforeseen scenarios, thereby significantly reducing false positives.
3. A right combination of hardware was used to ensure effectiveness of solution. Separate hardware components helped to (1) Allow for future expansion of sensors or storage (2) swapping of specific component as part of warranty and service, rather than the whole integrated unit (3) A separate speaker ensured a louder output in a noisy cabin.

THE RESULTS

Through continuous education and periodic safety briefings of the drivers, DHL was able to bring down the number of violations across categories over a period. Some of these incidents include –

• Driver camera blocked
  

• Driver Smoking
  

• No Seat Belt
  

• Wrong Way Driving
  

• Illegal overtaking
  

CUSTOMER TESTIMONIALS

“Our company takes safety of the driver very seriously. In case of an unfortunate incident that impacts the driver, a thorough investigation is carried out where each member involved in scheduling and facilitating the journey must submit details as part of the inquiry. The investigation report is to ensure that safety processes are strictly followed and if need be, altered, to avoid future lapses. The AI & video telematics solutions helped to not only alert the driver, but the videos also support driver’s perspective.” – [DHL Representative]
“With the launch of the solution we saw a spike in the driver violations to company policies and traffic rules, an insight that could not be otherwise validated. The violation videos were made part of the driver de-briefing after every trip completion. The videos also gave inputs into certain difficulties that the drivers faced, and changes were implemented towards better driving conditions.” – [DHL Representative]

CONCLUSION

Today, with close to 100 vehicles installed, DHL is now tracking driver behaviour; and top drivers are awarded. This also works as an incentive for others to drive better, which helps in their own safety and also maintenance of the vehicle.

OmniEyes has helped DHL with information that they had no visibility to. DHL now has insights to details about drivers, their habits, patterns, and related reasons, leading to help DHL solve for issues that may be impacting the driver’s work environment. Lack of evidence often leads to trust factor, and video recordings help in bridging the gap providing credible information.

In other markets too, AI & Video telematics solutions have had a significant impact in first identifying the events, and later reducing these events, benefiting all involved users. With the implementation at DHL being recent, the expectation is to replicate the same success once the project deployment is complete.

The post DHL first appeared on OmniEyes | 動見科技.

Background: Giant Group and YouBike
Giant Manufacturing Co. Ltd., commonly known as Giant, a global leader in bicycle manufacturing with extensive market presence, has long been a leader in sustainable urban mobility. As part of their mission to enhance eco-friendly transportation, they operate YouBike, Taiwan’s leading station-based bike-sharing service. With more than 7900 stations conveniently located near public transit hubs and city hotspots, YouBike has become an integral part of Taiwan’s urban commuting network. The annual YouBike rental numbers among the six metros, including Taipei, Taichung, and Kaohsiung have exceeded 160 million in 2024, highlighting its widespread adoption and significance.

To ensure a seamless user experience, YouBike manages a dedicated fleet responsible for redistributing bicycles across its stations. These trucks operate in busy city environments where traffic violations, particularly red light running, pose safety concerns. Recognizing the importance of safe and efficient fleet management, YouBike sought an advanced AI-driven solution to improve driver safety and operational oversight.

The Challenge: Enhancing Driver Safety and Compliance

With a fleet that continuously navigates high-traffic areas, YouBike faced several key challenges:

• Reducing overall traffic risks, including red light running, no stop at intersections, as well as real-time alerts to drivers to avoid accidents, to protect pedestrians, cyclists, and drivers alike.
• Improving fleet management through AI-driven insights and driver behavior monitoring.

In 2023, YouBike made a strategic decision to implement OmniEyes’ AI-powered fleet management solution, MyFleet. Over time, they expanded their fleet from nearly 200 to close to 300 trucks, all equipped with OmniEyes technology. This investment aimed to provide greater oversight of driver behavior, fostering a culture of safety and accountability.

The Solution: AI-Powered Monitoring and Driver Coaching

OmniEyes’ MyFleet system continuously monitors driver behavior in real time, detecting critical safety concerns such as red light running and failure to stop at intersections. It also provides real-time alerts to drivers, helping them avoid collisions with vehicles or pedestrians, ensuring a safer driving environment. The solution provides instant feedback and actionable insights, allowing fleet managers to identify high-risk driving behaviors.

Starting in April 2024, YouBike leveraged OmniEyes’ AI-generated insights to launch a structured driver coaching program, specifically targeting red light running and intersection compliance. This proactive approach equipped drivers with the necessary awareness and tools to adopt safer driving habits. This targeted intervention empowered drivers with real-time awareness and reinforced safe driving practices.

The Results: A Measurable Reduction in Traffic Violations

The impact of OmniEyes’ AI-powered intervention was both rapid and measurable, reinforcing YouBike’s commitment to continuous safety improvements. Key improvements observed include:

• A 53% reduction in red light running violations from April to December 2024.
• A 94% overall improvement in driver behavior after the implementation of coaching programs.
• Increased adherence to traffic regulations, contributing to safer roads and enhanced public confidence in YouBike’s commitment to safety.

Customer Testimonial: The YouBike Experience

“At YouBike, safety is paramount—not just for riders using YouBike but for everyone on the road. Implementing OmniEyes’ AI solution has transformed the way we manage our fleet. Since we began leveraging MyFleet, our drivers have become significantly more aware of their behaviors, leading to a notable drop in traffic violations.” – [YouBike Representative]
“The data-driven insights provided by OmniEyes allowed us to implement targeted coaching and behavioral corrections. The results speak for themselves, and we are thrilled with the improvements we’ve seen in fleet safety and compliance.” – [Fleet Manager, YouBike]

Conclusion: A Safer Future with AI-Driven Insights

By integrating OmniEyes’ MyFleet AI technology into their operations, YouBike has significantly enhanced driver safety, improved compliance, and contributed to safer roads across Taiwan. This case study underscores the power of AI-driven insights in fleet management, demonstrating how data-informed decision-making can drive impactful change.

With an ongoing commitment to safety and operational excellence, YouBike and OmniEyes continue to work together to create a safer, smarter urban mobility ecosystem for the future.

The post U-Bike first appeared on OmniEyes | 動見科技.

Unlike conventional all-in-one dashcams, OmniEyes’ modular AI Edge Gateway architecture supports multiple camera configurations — including road-facing and in-cabin views — as well as peripherals such as OBD dongle, thermometers, door sensors, alcohol breathalyzer, TPMS, panic buttons, and more. This flexibility enables seamless deployment across a wide range of vehicle types, from light commercial vans to heavy-duty trucks.

At the heart of the upgrade is the new Accessory Hub, an integrated expansion module that consolidates key components like the speaker, microphone, panic button and a USB hub with 4 ports. Positioned on the dashboard, it connects directly to road-facing and in-cabin cameras, while a redesigned OBD dongle supplies power for true plug-and-play installation.

Key benefits include:

• Up to 70% faster installation — setup time reduced from 60–90 minutes to just 20–30 minutes
• DIY-friendly — simplified wiring and configuration lowers adoption barriers
• Expandable design — add cameras or modules without replacing the main unit

The upgraded AI Edge Gateway now connects existing MDVR camera systems to the cloud for remote access and live streaming — no hardware replacement required. This backward-compatible approach opens new collaboration opportunities for system integrators and fleet partners.

“Our mission is to help fleets solve costly operational issues and liabilities that require tailored video solutions from all angles around a vehicle,” said Tyson King, CCO of OmniEyes. “Accessory Hub combined with MDVR capabilities in the AI Edge Gateway unlocks this potential — delivering customizable, yet simple AI video deployments cost-effectively.”

Mass production of the OmniEyes Accessory Hub will begin in Q1 2026, with rollouts planned across North America and Asia.

OmniEyes will also introduce its AI Edge Gateway + MDVR solution in partnership with Tadi Brothers, combining their onboard monitors and wireless cameras to create a plug-and-play, future-ready AI video telematics platform.

“As a proudly American-based company with over 20 years of experience delivering reliable 1080P wired and digital wireless camera systems, we are very excited to partner with OmniEyes. Together, we are introducing a truly unified fleet solution to market — combining our proven hardware with their advanced AI video telematics platform. This partnership provides fleets with an easier, more reliable way to upgrade existing vehicles into connected, intelligent systems that enhance safety, visibility, and operational efficiency without the complexity of traditional retrofits”, said Tamir, Owner of Tadi Brothers.

The post OmniEyes Unveils Next-Generation Modular Video Telematics Solution Featuring Accessory Hub and 8-Channel MDVR first appeared on OmniEyes | 動見科技.

Currently, both regular and commercial vehicles can enhance driving safety by extensively applying driving image equipment and technology through either pre-installation or post-installation. For example, regulations require larger commercial vehicles over 3.5 tons to be equipped with a driver assistance system with at least four cameras. These images can be displayed in real-time on dazzling screens for driver reference, stored in in-vehicle devices like DVRs for future accident investigation, or even streamed live to the cloud for real-time monitoring by management or driver live streaming. They can also be stored in the cloud for big data analysis using various AI technologies. Consequently, the value of images is rising rapidly, with a promising outlook.

To conclude, more money is always better. Just as the moon waxes and wanes, and life is full of uncertainties, who knows what the future holds? However, poverty is said to limit imagination, and truly wealthy people worry about being so poor that all they have is money.

Whether through government regulations or existing fleet management technologies, driving images are often viewed as a form of currency—the more, the better. Just as the moon waxes and wanes, and vehicles face uncertainties, who knows what might happen in the future? It’s best to store everything, both in the vehicle and in the cloud. As for how much is stored, how much is used, and whether it becomes something people like, that can be discussed later.

Let me temporarily put on the hat of a National Taiwan University professor and propose a method to turn images into something as desirable as money: Once images are stored in the cloud, they can undergo various image recognition or AI computations using the cloud’s powerful computing capabilities. Without the limitation of computing resources, the most accurate models and highest quality images can be utilized. If that’s not attractive enough, images uploaded from different vehicles can be integrated and analyzed to build various smart city applications, turning these invaluable images into something golden.

As for how much money needs to be spent initially to generate more money, a professor only needs to focus on imaginative research without worrying about costs. However, as the CEO of a startup, I can clearly tell you that customers won’t spend $1,000 to generate $100 worth of gold. The costs of the three major cloud platforms are available online and are more transparent than taxi fares. We spend a considerable amount on AWS every month, and the service is decent, but the proportion of AI computing costs is zero—not because it’s free, but because it’s simply unused.

The post AI Fleet Management Series Part 2: Is Image All-Powerful? first appeared on OmniEyes | 動見科技.

Data shows that the risk level for commercial fleets or professional drivers is several times higher than that of general drivers (this can be objectively compared through accident rates and insurance premiums). Let me first clarify that this is not entirely because professional drivers are worse. If you drive on the road for 8-10 hours a day, need to park by the roadside dozens of times, get in and out of the vehicle to load or unload, or let passengers on and off, and on top of that, deal with complaints from bosses or customers about late deliveries, your record might not look so good either.

However, the data is clear: poor or even dangerous driving behavior is a long-standing issue that any commercial fleet must face and resolve. Where there is demand, there will be supply—this is an unchanging law of market economics. As a result, various products and technologies for detecting and monitoring driving behavior flood the market, many of which claim to be backed by scientific evidence. Just like various health supplements, they not only always receive certification from some unheard-of authoritative British institution but also often claim to be the result of years of effort by doctors or professors from National Taiwan University.

The effectiveness of many fleet management technologies is as inscrutable as the medical effects of these supplements. While data itself does not lie, those who select and interpret the data might. This time, let’s explore two indicators that are often valued in both the Taiwanese and Western markets: Harsh braking and Harsh acceleration.

From the perspective of the average person, harsh braking or acceleration is indeed a negative indicator of driving behavior. For commercial fleets, it may lead to frequent accidents or unnecessary fuel consumption, making it a good basis for fleet management. Products that obtain these two data points using OBD, CAN, or gyro/accelerometer are as numerous as fish in the sea.

To conclude: if you’re hesitating about whether to bring an umbrella, just look at the probability of rain, not the height or thickness of the clouds. While understanding cloud information is helpful, looking at the probability of rain is more direct and effective.

Back to the topic, when do people usually brake harshly? Aside from professional tofu delivery drivers like Takumi Fujiwara, it’s usually when following too closely (poor driving), being aggressive or being aggressed upon (the former is poor driving, the latter is helplessness), or encountering unexpected obstacles on the road (emergency defensive driving). Can harsh braking data distinguish between poor, helpless, or emergency defensive driving? Maybe, but it’s like observing cloud height and thickness—not direct or effective. Is there a more direct and effective method? Yes—direct video analysis.

When do people usually accelerate harshly? Typically, it’s to beat a red or yellow light (poor driving), aggressive driving (poor driving), maintaining speed uphill (normal driving), or merging into fast traffic (normal driving). Can harsh acceleration data distinguish between poor and normal driving? Maybe, but it’s not direct or effective. Is there a more direct and effective method? Yes—direct video analysis.

You might say that video analysis can’t analyze fuel consumption, right? From my understanding of various fuel consumption analysis methods, I’ve never seen one using video, because it’s not direct or effective. But can harsh braking or acceleration analyze fuel consumption? Maybe, but isn’t it simpler and more direct to use the total mileage/total fuel consumption figures for a single trip, day, week, or month? Isn’t this elaborate process similar to deciding whether to bring an umbrella by looking at cloud height or thickness?

The post AI Fleet Management Series Part 1: Driving Behavior vs. Vehicle Condition first appeared on OmniEyes | 動見科技.