5 min read

A Highway Safety Milestone: MicroVision’s June Test Track Update

New hardware. New scenarios. New evidence that MicroVision is leading the pack when it comes to helping OEMs deliver high-speed safety features.

This June, our team was back at the track, completing another important milestone on our path to securing OEM partnerships. And while the track itself might look familiar to the testing we demonstrated in March, we have many exciting updates to share.

New Hardware. MicroVision MAVIN™ DR

The recent track testing put the new MicroVision MAVIN DR lidar hardware to the test in real-life driving scenarios. Now I know you might be thinking, ‘this doesn’t look like the MAVIN DR’ that we announced in June—with the white shell and new, more organic design. Rest assured, the lidar hardware mounted atop our test vehicle packs the same dynamic view lidar combining short-, medium-, and long-range sensing; ultra-high resolution; low latency; and quieter MEMs into an ultra-low profile form factor. It just came in black 😉 Unlike March, when our test car was outfitted with separate short and medium-long range lidar sensors, this time, our test vehicle was only equipped with one sensor – MAVIN DR – combining all fields of view into one.

New Scenarios – Now More Complex and Realistic

Building on our confidence from the March testing and equipped with MAVIN DR, our goal was to raise the bar this time with more complex real-life highway driving scenarios. In speaking with leading OEMs around the world, we understand better than most the complex, high-speed ADAS safety problems that automakers are focused on. We wanted to demonstrate that MicroVision is taking a superior approach and is surpassing the already high standards set by OEMs to deliver highway speed safety features. That’s why this time around, we added a number of new elements to our tests that made them more challenging and realistic.

Changing lighting conditions. Tunnels. Moving vehicles. Oh my…

In March, we demonstrated how our hardware excels in changing lighting conditions. This time, we’re adding new elements to our tunnel scenario to reflect the challenges drivers face every day.

Changing lighting conditions often hide threats from obstacles and other vehicles. Just as our eyes take time to adjust, camera-based systems are slow to react as well. If another vehicle in the tunnel was to brake suddenly, or change lanes, or if there was an obstruction around the corner in the tunnel, drivers would be left with just milliseconds to take corrective action and avoid an accident. That’s what we demonstrate in this new tunnel scenario.

MicroVision MAVIN is able to maintain a high-resolution view of the road ahead as the test car enters the tunnel. We see the other vehicles traveling alongside it. When one vehicle swerves suddenly to avoid an obstacle around the bend, we pick it up instantly. Our low latency, high-resolution point cloud and data are fully pipelined into the OEMs ADAS without additional object classification. This translates to the vehicle being able to respond faster and make decisions that will keep passengers safe.

Microvision's new high speed automotive lidar MAVIN approaching a tunnel

MicroVision Tunnel

Vehicle tracking at high speeds

Highways present challenging driving conditions for even the most experienced drivers. There’s an incredible amount of data to process – signs, lane markings, and other vehicles that are also moving. With all this information coming at you so quickly, dangerous scenarios can unfold in fractions of a second. Because highways represent a dynamic environment where everything is moving, vision-based sensors can often and easily get confused. When a vehicle becomes obscured by another car on the road as a result of changing lanes or being passed by another car, these vision-based sensors lose sight of these objects.

We’ve set up this scenario to show how MicroVision MAVIN offers a superior approach. As a pack of vehicles makes its way along a three-lane highway, the vehicles are crisscrossing back and forth in front of each other. By combining dynamic range with low latency and an ultra-high resolution point cloud, MicroVision MAVIN DR can maintain a clear picture of the drivable and non-drivable road at all times.

This scenario also demonstrates MicroVision’s better approach to understanding the velocity of other vehicles on the road.  Our system looks at an object across multiple frames. Once an object like another vehicle is being tracked, we can determine the velocity of that object instantaneously and continue to track its velocity consistently—even as it moves across the road. MicroVision MAVIN captures two components of velocity—axial and radial. This allows us to look at velocity holistically and understand the vector that other vehicles are traveling on. This vector input is a critical component that ADAS systems require to determine whether another vehicle on the road is moving sideways – as would be the case if it was about to change lanes and cut you off.

Other lidar systems that claim to offer ‘instant velocity’ simply aren’t up to par. Since they’re doppler-based, they miss an entire dimension to velocity and, as a result, are less useful to OEMs. This disparity between approaches is not lost on OEMs. MicroVision being able to demonstrate this stronger understanding of how velocity needs to be done correctly is resonating with OEM partners in our conversations.

MicroVision's high speed automotive lidar MAVIN approaching a pack of vehicles

MicroVision Swarm

A watershed moment for highway safety

We believe that this recent round of testing is a watershed moment for MicroVision and the lidar landscape. We’ve been able to demonstrate how MicroVision’s approach to high-speed safety surpasses the already high standards set by OEMs. MicroVision MAVIN DR highlights what’s possible when it comes to enabling OEM ADAS capabilities to address complex highway safety challenges. And we continue to meet our go-to-market milestones and gather important ground truth data that will allow us to continue refining our hardware and software.

What’s next? We’re on pace to achieve Class 1 in Q3, which will pave the way for sample sales to OEM and other customers.