Game-changing LiDAR sensors for safer self-driving cars
Photo credit: metamorworks / Shutterstock.com
A US-based self-driving car developer, partially owned by Ford Motors and Volkswagen, could shift the focus to the burgeoning and turbulent LiDAR sensor industry. Argo AI developed the new LiDAR sensor for autonomous cars in-house following a recent acquisition.
New LiDAR sensors for autonomous cars
The new sensor uses LiDAR technology to create a 3D digital map of its physical environment. Argo’s promising contribution to LiDAR technology was developed for the autonomous vehicles (AVs) the company makes.
The sensor is more accurate than ever and has reportedly enabled safe self-propelled speeds of 100 km / h. This means that future Argo AVs equipped with it will be able to drive safely on highways, enabling the company to address the logistics and freight industries by developing self-driving trucks.
Only a handful of LiDAR sensors are currently available for self-driving cars. They are usually developed and manufactured by engineering firms with a product or technology. There are competing LiDAR technologies as well, but Argo’s support from major auto companies will ensure that its approach gets a good share of the market interest.
The AV industry has yet to reach consensus on sensor technologies in general, let alone various LiDAR systems. However, Argo’s decision to develop its own LiDAR technology for the AV equipment it manufactures also means a potential shift in priorities for the entire AV industry.
What is LiDAR?
LiDAR stands for Light Detection and Ranging and is a type of sensor technology that is often found in AVs. With LiDAR sensors, AVs can see around them and create a digital map of the physical environment in real time. This in turn enables the on-board processor to make decisions and drive the vehicle.
LiDAR sensors often fire focused beams of light and then detect any light that is reflected back towards the sensor. If the light is reflected back towards the sensor, it has hit an object some distance away. LiDAR sensors can tell how far away that point is on that object by measuring the time it takes between leaving the light and returning to the device.
A LiDAR sensor combines multiple beams of light that are sent in different directions and repeated many times to produce millions of points of 3D geospatial data. This data can be represented visually as a network. The more points the sensor can record, the higher the resolution of this digital 3D network. In AV applications, higher resolution means better accuracy – and safer vehicles.
The benefits of LiDAR sensors for autonomous cars – as argued by proponents of the technology – are the potential for a wide range of operation, speed, and accuracy. LiDAR is particularly accurate in precipitation and poor lighting conditions. Under these adverse conditions, other AV sensor technologies such as cameras also have difficulty performing.
Big changes in the LiDAR sector
There are currently two main types of LiDAR sensors under development and in prototype use for AVs: mechanical and solid-state sensors. Mechanical sensors like Argo are a more established technology with a much higher level of accuracy. Solid-state sensors are the emerging LiDAR technology. Although they are cheaper and lighter, they are much less fragile.
Mechanical sensors are usually designed to sit on top of AVs and look like a large box or siren. These devices mechanically rotate the light source in a circle and can thus capture spatial data 360 degrees around the AV. Because of the mechanisms that need to be developed for this type of sensor, they are typically much more expensive to manufacture.
Solid-state sensors do not move their light source at all. Instead, they arrange light sources to get the widest possible view of up to 180 degrees. Solid state sensors are much cheaper to manufacture because they have no moving parts. They are also lighter and can be made in a much smaller shape.
Ford, which shares an equal 80% stake in Argo alongside Volkswagen, had previously invested in solid-state LiDAR sensor technology. Velodyne – with $ 75 million in funding from Ford – recently announced plans to launch a solid-state LiDAR sensor with a price target of just $ 500. Combined with the small form factor that solid state offers, the sensor could be well suited for home AV production.
The significantly higher accuracy that mechanical sensors can offer, however, has significantly exceeded the advantages of LiDAR for solids for Argo. The company’s relatively unusual approach to in-house development continues to signal its support for the older technology.
It also signals a possible demolition from the perspective of the broader AV industry, inevitably dictated by its largest and most outspoken consumer maker – Tesla. Tesla notoriously rejects LiDAR sensor technology instead of using cameras as the basis for their semi-autonomous and autonomous driving systems (autopilot and full self-driving).
Developers like Argo – with investments from major automakers – can win the conversation with the superior performance and accuracy of LiDAR. Argo’s own contribution to LiDAR technology further strengthens the argument.
References and further reading
Pilkington, Ben (2020) Velodyray H800 from Velodyne LiDAR: Self-driving cars with improved vision. AZO sensors. [Online] https://www.azosensors.com/article.aspx?ArticleID=2107.
Vousden, Mike (2021) New Argo LiDAR sensor opens the highway for self-driving cars. Car only. [Online] https://www.just-auto.com/analysis/new-argo-lidar-sensor-opens-up-the-highway-to-self-driving-cars_id201782.aspx.
