Why Lidar Will Advance Safety and Efficiency for the Rail Industry: A Conversation with Akram Benmbarek, VP of Industrial Markets & Strategic Initiatives
If current trends continue, both passenger and freight rail activity is expected to more than double by the year 2050. While this growth indicates social and economic progress, it also leads to a greater need for technology that ensures both safety and efficiency.
To meet this challenge, AEye has developed 4Sight™ lidar for rail, a software-definable solution that delivers the industry’s longest detection range and can deploy high resolution Regions of Interest (ROIs) for small object detection and short range applications. We sat down with AEye’s VP of Industrial Markets & Strategic Initiatives, Akram Benmbarek, to learn about the advantages of AEye’s software-definable lidar for commercial and freight rail applications.
What’s the advantage of using lidar for rail?
AB: Lidar bridges the gap where cameras and other sensors fall short, and that mostly comes down to providing high quality data. Safety and autonomy rely on this data, whether you’re talking about rail, ADAS, or ITS applications. You can’t have reliable control over autonomous systems and take necessary evasive action unless you have accurate information. Lidar uniquely provides the highest quality data for two reasons:
First, lidar detects in all lighting and weather conditions. If you use a camera at night, you’re not going to see a hazardous object in the distance. Similarly, radar has limitations in detecting lateral movement, and does not offer the resolution necessary for the system to make an informed decision.
Second, lidar gives you an accurate 3D representation of the world. This is an upgrade over cameras, which only provide data in 2D. In those 2D scenarios, a camera collapses objects in the horizon to a single pixel, so you don’t know exactly what that object is, its size, or the direction it’s heading. The object could be a person or a vehicle or something else entirely. Alternatively, lidar provides 3D data with precise measurement of the object’s dimensions, distance, and velocity, which increases the accuracy of classification. This precision is essential for safety, whether for rail or for other transportation applications.
What does AEye do differently that ensures better nighttime vision and higher resolution?
AB: Lidar in general has great advantages, but not all lidar sensors are created equally, and there are differentiation factors that set AEye’s 4Sight lidar apart from the pack. One is that AEye’s lidar uses a 1550 nanometer bistatic architecture that allows it to leverage its generous photon budget to achieve long-range detection, even in adverse lighting or weather conditions.
More importantly, AEye’s is the first solid-state intelligent lidar. And what I mean by that is our lidar has embedded processing capabilities and built-in intelligence that allows it to configure and optimize a scan pattern to any given use case, and switch from one scanning configuration to another in real-time. We can trade off resolution for range, giving enough power to detect all the way up to a thousand meters.
Let’s say I come across a situation where a sensor is mounted on a train approaching a station. As the train approaches within 200 yards of the station, it no longer requires long-range detection. Our sensor allows it to trade that long-range capability for a higher resolution, near-range scanning configuration where the train is able to detect and track passengers in the station. With AEye’s intelligent lidar, railways are able to use the same lidar sensor, adaptable via software, for numerous applications and use cases.
Can you explain the changing of scan patterns from one to the next?
AB: This is all automated and programmable. We design and optimize scan patterns for different use cases to provide the most accurate 3D spatial data possible for any given application. In the case we just discussed, the train is slowing down as it approaches the station. I can program the sensor or the pattern to be triggered based on an IMU or a GNSS, so as my speed decreases to a certain level, a new scan pattern is triggered. The train doesn’t need to see too far in the distance because it’s going to come to a complete stop soon, so the intent here is trading off long range capability for higher resolution.
When it comes to rail, is longer range, a wider field of view, or resolution more important? Or are they on an equal level?
AB: The requirements will be different for each use case, but by trading off one of those parameters, you’re able to maximize the others. So as a train approaches a platform, you want to have a wider field of view in order to see it. As we widen the field of view, we lower the range but increase the resolution and deliver a high-quality frame rate.
What are the main applications in the rail industry stakeholders are pursuing?
AB: Number one is asset and track inspection. Right now, it’s onerous to inspect tracks if you use special vehicles mounted on the track or do it on a drone because you need multiple products specialized for individual use cases. But with AEye’s 4Sight lidar, you can easily mount a sensor on a revenue train and use it for asset inspection and safety, simultaneously.
Second is shuting, which is moving train cars or locomotives from one track to another. Conventional lidar sensors do not deliver a range of distance capabilities, and that creates a safety hazard. You need that adaptability in train yards especially, where moving cars from one track to the next can cause accidents and even fatalities. AEye’s lidar increases safety for these operations with range capabilities around 150 to 200 meters.
Are there any other pressing issues in rail that AEye’s lidar solves?
AB: In the rail industry, sensors are typically mounted on trains, and the fact that AEye’s lidar sensor is solid-state and MEMs-based is an advantage. We have the smallest MEMS in the industry, which allows our sensor to sustain high shock and vibration in any environment – essential for rail. And, of course, this is different from conventional lidar sensors which have mechanical articulation that moves the laser, making it too unstable to meet customer reliability requirements.
How does AEye work with partners to ensure its solutions are compatible within the software stack – with existing and future infrastructure and vendor technology?
AB: That’s pretty easy because the sensor data is natively simple to consume. At the end of the day, it’s point cloud data that gets processed through a perception algorithm. This perception algorithm puts out very simple data, delivered in universal formats. So really the advantage of our lidar sensor is that it’s simple to integrate in any new or existing application.
What should rail customers remember about AEye?
AB: AEye’s 4Sight lidar for rail enables complex use cases with long-range requirements. Rail is a vertical where AEye is second to none because our long-range capability has been proven time and again, and our ability to optimize and configure scan patterns for different use cases is critical for any moving vehicle or transportation application. AEye is very much poised to win this race.