- #movingpeople
- Posts
- #movingpeople-meets Eye-Net
Eye-Net is a smart collision-prevention system. High level - what do you do?
Eye-Net was founded to protect vulnerable road users (VRUs) by enabling early warnings for both cars and people. Today, ADAS systems exist almost entirely inside the vehicle - everything is car-centric. Pedestrians, cyclists and scooter riders don’t really have anything protecting them.
What we’re trying to do is give an early alert that a collision is about to happen to both sides, not just the driver, but also the person outside the car.
We focus on what we call non-line-of-sight scenarios. For example, someone crossing the road from behind parked cars, or a cyclist or scooter coming out from behind a bus or a building. No matter what sensors you use - camera, radar, LiDAR - these situations are detected too late. It’s always the last moment.
Our system works via the mobile phone, using GPS and cellular networks. We track road users in real time, and once we determine with high certainty that two objects - whether pedestrians, cyclists, scooter riders, or vehicle drivers - are on a collision path, we send a warning. To both sides.
Two big challenges immediately come to mind: latency and accuracy. How do you deal with those?
We’ve invested a lot of effort there. Today we have five patents dealing exactly with these two problems, and we’re in the process of filing more. In terms of latency, we’re below 100 milliseconds, which, for example, is similar to latency in remote driven vehicles.
On the positioning side, in one of the recent pilots, the deviation of accuracy reached was 20cm. These two numbers – latency and accuracy – are really what allow the system to work.
And we’re doing all of this without any dedicated hardware. Almost everyone else in this space relies on hardware. We’re a pure software solution, and that’s what sets us apart.
#movingpeople is a part of Mobility Business - a consultancy dedicated to "All Things Mobility", focused on growth.
You’re running continuously in the background. What does that do to the phone’s battery?
That was a big concern for us from day one. We designed a very specific algorithm to manage this. The CPU usage is extremely low, and the impact on the battery is minimal.
We always assume that we’re being hosted inside someone else’s app, so we can’t afford to interfere with their performance or user experience. The system runs quietly in the background, and in practice the user doesn’t really notice it’s there.
You have three different products – Protect, Sense and View. How do they differ?
Eye-Net Sense is designed for the automotive side. It’s integration-ready and this is the product we offer to Tier-1 suppliers. The core technology is the same, but Eye-Net Sense is meant to sit inside the vehicle - typically on the telematics control unit (TCU) or electronic control unit (ECU). We’re seeing traction here, mainly driven by V2X regulation around ADAS and autonomous systems. You see the industry moving in that direction. The Qualcomm–Autotalks acquisition is a good example.
Eye-Net Protect is the smartphone product. It’s for cyclists, pedestrians, scooter riders - but also for drivers if they’re using Eye-Net on the phone. It adapts itself automatically to the user profile based on how the user is moving, so the same app works for different types of users.
Eye-Net View is our data product. It’s aimed at cities and governments, but also at industry players like ride-hailing companies. It delivers analysis of accident and near-miss hot-spots, traffic patterns and event validation.
That helps understand your target customers, but walk us through those.
We’re targeting both B2G and B2B customers.
On the B2B side, we work with third-party mobility apps - e.g. ride-hailing, micromobility, MaaS and super-apps. We integrate through an SDK that sits inside the host app. If we’re embedded in more than one app on the same phone, we’re aware of that, so we avoid creating false collision scenarios and we make sure the user doesn’t get duplicate alerts.
We also work with Tier-1 suppliers, so our software can be integrated directly into vehicles. In that case, we don’t need to rely on the driver’s phone – the car itself comes ready with our early-warning technology. A good example is our work with Aumovio (formerly Continental).
On the B2G / smart-city side, we’re currently running a pilot with Software République in France. The project in Bordeaux involves partners like Keolis, Orange as the telecom provider, TBM, the local public transport operator, and others. The focus there is on protecting VRUs. We’re still in the middle of the pilot, and we expect results by early 2026.
Geographically, our focus is quite broad – Europe, the US, and Southeast Asia.
Let’s go back to the solution itself. How does the user actually experience it?
There are a few options, and the host app decides which ones to use.
It can be sound – a very distinct beep.
It can be visual, for example, the flashlight turning on or a screen alert.
It can be a vibration.
And it can also be a combination of all of these at the same time.
We also define different alerts for different user types - pedestrians, cyclists, drivers,, and so on. The system determines the type of user by analysing speed, manoeuvring, location and direction of travel. Based on that, it triggers the relevant type of alert for the specific situation.
And how much warning time does the user actually get?
It depends on the user category. A driver will usually get around three seconds, while a pedestrian is closer to two seconds. These timings are based on statistical analysis we’ve done, to give people enough time to stop or change behaviour and avoid the accident.
It’s not a fixed number either – it can be adjusted depending on the environment and the specific use case.
What is your business model?
Our main focus today is a licence-fee model. We also have revenue sharing schemes.
What should people reach out to you about?
Any B2B partner or client reading this - ride-hailing, MaaS, micromobility platforms - is very welcome to reach out. Insurance companies are also a strong fit. Fewer accidents mean fewer claims, and that value can be passed on to users in the form of lower premiums, which is especially relevant given how motor insurance prices have increased in recent years. On top of that, our technology also helps with fraud detection. In ride-hailing, for example, there are cases where riders submit false accident claims. With our data, those cases can be identified much more easily.
We also want to speak with municipalities and governments, particularly in areas with high accident rates, to see how this can be implemented at the city level. And lastly, we’re also interested in working with local safety organisations, especially those focused on vulnerable road users.
We’ll be presenting at CES in Las Vegas in January, so if you’re there, please come by and see us.
Tell me about your journey - you are a three time founder yourself.
Yes. The first, Atlas, was in entertainment – automation technology that significantly reduced video editing and production time, which opened up low-budget markets. I sold my share after two years.
The second was in smart homes: a mobile-controlled door lock that still kept the physical key option, which solved a real adoption barrier. That company was acquired by an established door-lock manufacturer.
The third was smart parking - guiding drivers directly to available spaces using Bluetooth beacons and light camera use, without expensive infrastructure. That one didn’t scale, and I eventually shut it down.
After that I joined Intel, working on M&A around the Mobileye acquisition, and then spent six years at Foresight Automotive, Eye-Net’s sister company, focused on stereo 3D vision for object detection. Moving to Eye-Net was a natural step, given the overlap in technology and the focus on road safety.
So, as a three-time founder, what advice would you give other entrepreneurs?
You will always hit obstacles – that’s part of the job. When you do, don’t try to go straight through the wall. Most of the time, that doesn’t work. Find a way around it, over it, or from the side.
That’s the advantage startups have. We’re nimble, we can think differently, and we can move faster than large organisations. Use that flexibility. Be ready to adapt, change direction when needed, and keep pushing forward without losing momentum.
#movingpeople is a weekly mobility & delivery newsletter - from Ride-Hailing & Robotaxi through Remote-Driving & Delivery Drones to e-bikes and Autonomous Freight & Logistics.
Thank you for reading. If you like what you're reading, please share it with your friends and colleagues so they can benefit from it too.