How LiDAR Technology is ensuring Self-driving Cars’ Safety

Autonomous cars are disrupting the automotive industry and may not be as far as we think of becoming transportation’s future. Self-driving cars have hit the streets in a few states but for testing purposes in restricted areas. Though one cannot purchase a fully autonomous vehicle (AV) today, semi-autonomous technology is available in vehicles for some time. Creating an entirely AV will require incorporation of semi-autonomous technology silos. In this scenario, Vision-enabled light detection and ranging (LiDAR) technology is a key component in bringing safe self-driving vehicles to life.

LiDAR is the latest development in surveying technology, advancing on the shoulders of its predecessors, sonar and radar. Rather than using sound or radio waves to scan its environment, LiDAR utilises laser light pulses. By doing so, lidar systems can navigate surroundings at the speed of light. Its versatility in direct air and in the vacuum of space enables lidar to operate on a short-wave, near-infrared optical signal delivering finer scan accuracy.

Significance of LiDAR in AVs

Lidar allows AVs to make calculated decisions with its ability to identify objects in its immediate environment. It can be considered as a vehicle’s ‘set of eyes’ and the most crucial component in making AVs a reality. Automakers are leveraging lidar technology to develop safe self-driving cars.

Lidar systems can see things beyond human capabilities. Suppose if one’s eyesight provided continuous 360-degree visibility or one’s depth perception was always correct and the individual never had to guess stopping distances between one’s car and those cars in from of the individual. Lidar makes it possible, and it is a technology of precision in collecting data and computing accurate distances that promote AVs’ safety.

But it throws two significant challenges:

High Computing Requirement:  Rich lidar data processing makes the technology much more expansive than its counter-parts camera and radar that have been in the automotive industry remarkably longer

Evolving Designs: There are different types of lidar from solid-state scanning, solid-state flash, rotating SMSE, FMCW, and more

Technology manufacturing company like Xilinx addresses above-mentioned challenges. Its powerful DSP capabilities coupled with flexible I/O configurations and programmable logic are a good fit for the high compute lidar makers. Additionally, its devices consist of programmable hardware that can adapt to any lidar sensor configuration, making it ideal for varying and evolving designs. There is not any clear ASSP/ASIC device architecture as lidar technology is comparatively new, and a common approach has not been embraced in the ADAS/AD market.

To meet the high computer and evolving designs’ needs for lidar, some technological companies’ solutions are also helpful for addressing cost and power issues. FPGAs allow real hardware-based processing pipelines for various sensor RX channels. It allows simultaneous and independent RX channel processing with differing objects. Additionally, it enables incorporated hardware stimulation for post-detection processing, i.e., point cloud generation and grid navigation and ideal partitioning between sensor software and integrated hardware acceleration functions using the high bandwidth connectivity from the processing system and programmable logic.

The integrated solution that FPGAs enable helps to cut down the cost. Also, parallel hardware processing reduces the requirement for clock speed, minimising power. This solution also offers unique opportunities to update sensor software and hardware re-programmability.