Jimi IoT on event-based video telematics: alerts, buffers and upload rules

    Vlad Tsveklinskiy
    AuthorVlad Tsveklinskiy
    January 29, 2026
    Two men discuss during 'Telematics Talks #10' against a bright geometric background.

    Video telematics has changed in a big way over the last few years. It is no longer a “nice to have” add-on for a small set of fleets. For many operators, it has become a practical tool that improves safety, reduces cost, and helps teams make decisions with real evidence instead of assumptions. The catch is that video only delivers value when it is designed for reality: patchy cellular coverage, expensive data, and limited time for people to review footage.

    In the latest episode of Telematics Talks, Vlad Tsveklinskiy from Navixy speaks with Oscar Zhang, General Manager of the DVR Business Unit at Jimi IoT. Jimi IoT is behind multi-camera AI dash cams used by fleets across more than 180 countries. The conversation focuses on what fleets and telematics providers ask about most often: how to capture the right clips, how to set buffers, and how to build upload rules that keep costs predictable while keeping evidence reliable.

    You can watch the whole episode on YouTube or listen to it here:

    Why fleets are shifting to event-based video telematics

    Continuous video sounds attractive, but it breaks down quickly in the field. When a truck goes offline, continuous streaming still misses key moments. When it does work, it can overwhelm teams with hours of footage that nobody has time to watch. The result is a system that costs a lot and still fails to deliver what matters most when an incident happens. That is why many fleets are shifting toward event-based video telematics, where the device detects important moments at the edge and uploads the clips that matter, with enough context to make them useful.

    Livestreaming vs event-based uploads in fleet video telematics

    One of the most common questions in video telematics is whether livestreaming is actually worth it. Oscar argues that livestreaming can make sense in specific situations, especially for short-haul operations carrying high-value goods where real-time visibility justifies the cost. For most logistics fleets, though, continuous streaming tends to be too expensive and often wasteful. It generates large volumes of data without delivering enough day-to-day operational value. In those cases, event-based systems provide a better balance because they focus on what matters and avoid pushing everything over the network.

    Fleet video telematics pain points: cost, safety, and compliance

    When it comes to the core challenges fleets face, Oscar groups them into three areas. The first is operational cost, which includes data bills and the financial risk of missing a critical video event that could affect liability and payouts. The second is safety and security, where reliable video evidence supports coaching and helps protect vehicles, drivers, and cargo. The third is regulatory compliance, which is becoming more important as new rules appear across different countries and regions. Noncompliance is not only about penalties. It can also limit a fleet’s ability to operate.

    Upload rules for video telematics: LTE for urgent alerts, Wi-Fi for the rest

    The real value of event-based video telematics shows up in execution. Instead of streaming everything, fleets are increasingly using edge AI to detect important moments and send only those clips, while maintaining a buffer so each clip includes context from before and after the event. Done correctly, this creates a workable formula: clearer evidence, predictable data costs, and less time wasted sorting through noise. The real challenge is configuration. Teams need to choose the right alerts, tune them carefully, and reduce false alarms so the signal stays high.

    A key operational principle is prioritization. Not every event needs to be sent immediately, and not every event should consume cellular data. Higher-priority alerts should reach the server as quickly as possible. Lower-priority items can be held and uploaded later, often when Wi-Fi becomes available. Clear rules help fleets avoid two common failure modes: sending too much and overspending, or sending too little and missing the evidence that matters.

    High-priority alerts in video telematics: SOS, collision, tamper, and jamming

    Oscar gives examples of what belongs in the high-priority category. A hardware SOS button is one of them. If a driver triggers it, the operations team needs to know right away, and it may justify turning on a live view for immediate response. Collision detection is another clear example, especially if there is potential injury and the team must react quickly. He also highlights security-related alerts that are often overlooked early on, such as camera blockage, tampering, and jamming detection. These events can indicate interference, which is exactly the moment when evidence becomes most important.

    Best buffer settings for event-based video: how much context you need

    A short clip after an incident rarely answers the real question of why it happened. Oscar shares a practical default used in many deployments: roughly seven seconds before the event and eight seconds after, which produces a 15-second clip. Fleets can extend this to 30 or 60 seconds if needed, and platforms can also support requesting longer footage after the fact, such as a one- to three-minute segment of normal recording. The key is to start with a buffer that covers most cases, then give teams a way to pull more context when it is truly needed.

    Reducing false alarms with smart alert prioritization

    Another principle that makes the event-based approach scalable is treating alerts as patterns, not just single moments. A single AI detection can be misleading. A driver may glance away once, or a sensor may produce a false trigger. But when the same event repeats multiple times within a short period, the risk profile changes. Systems can escalate those events and treat them as high priority. This logic reduces false alarms while still ensuring that real risk is handled quickly, and it supports more focused driver coaching.

    How to manage video review workflows for fleet operations teams

    Another cost fleets often underestimate is human review time. Even if data costs are controlled, pulling too many clips into the platform can overwhelm the people responsible for monitoring and coaching. Oscar suggests mirroring the device-side priority model on the platform side. High-priority events should be surfaced first for immediate review. Lower-priority events can be reviewed later as part of daily or weekly reporting. In practice, this is what keeps video workflows sustainable.

    Getting started with video telematics alerts without overwhelming teams

    For teams that are just starting, Oscar recommends avoiding the temptation to enable everything at once. A better approach is to begin with core driver safety alerts such as speeding, ADAS, and DMS events, then tune thresholds based on context like speed and driving conditions. A behavior that might be low risk at very low speed becomes far more dangerous on a highway, and alert logic should reflect that. He also returns to escalation based on repeated events, which can create a stronger safety outcome without drowning teams in notifications.

    Cross-border SIM and roaming strategies for video telematics fleets

    Connectivity is another major theme, especially for fleets operating across borders. Oscar discusses a growing preference for SIM solutions that reduce reliance on expensive roaming. Some providers offer multi-IMSI SIM cards that can switch identities across countries and connect locally, improving stability and reducing long-term cost. Even if the SIM itself costs more up front, it can be cheaper over time than paying roaming fees on every trip.

    What happens when vehicles go offline: smart retry and upload priorities

    Offline behavior is treated as a normal scenario, not an edge case. Trucks will lose coverage. The question is what happens next. Oscar’s recommendation is that when connectivity returns, the device should resume sending real-time data first, then begin uploading historical clips in the background. Security events like jamming should be treated differently, because they may indicate a threat rather than a simple coverage gap. In those cases, the system should protect the footage from being overwritten and upload it as soon as possible. The conversation draws a clear distinction between a vehicle going through a tunnel and a vehicle being jammed. Both look like “no connection,” but they imply very different operational responses.

    Cargo security with video telematics: camera layouts that work in real operations

    Cargo protection is a natural extension of the event-based approach, and it is often where video delivers the most tangible value. Many fleets start with a road-facing camera, but cargo losses often happen near doors, in depots after hours, or during stops where tampering can occur. Effective protection is less about adding cameras everywhere and more about placing them where they cover high-risk zones. Oscar describes a practical baseline layout that includes road-facing and driver-facing cameras for safety, optional side coverage for blind spots, and rear or cargo-area views that capture door activity. For many fleets, a single camera aimed at the cargo bay door is enough to support monitoring and investigation, while higher-value operations may justify broader coverage.

    Key takeaways: building a scalable event-based video telematics playbook

    The broader takeaway is that event-based video telematics is not just a feature. It is a system of rules. It relies on detection at the edge, clear priorities for uploads, sensible buffers for context, and workflows that respect both data budgets and human attention. When those parts work together, fleets can capture the evidence they need, control costs, and build processes that scale. That is when video stops being a heavy expense and becomes a tool that improves operations day after day.

    Contact sales to launch video telematics with Navixy

    If you are a telematics service provider or fleet operator looking to add event-based video telematics to your offering, Navixy can help you bring it to market with a practical deployment model and partner-ready integrations. Contact sales today to discuss your use case and get a demo.

    Share article