Autonomous vehicles will have the ability to visually lock on to objects and analyze them the same way human eyes do.
RoboSense has released a new LIDAR driving system for mass-produced autonomous self-driving vehicles. Using solid-state hardware and a deep learning sensing algorithm, the RS-IPLS Intelligent Perception LiDAR system features real-time data pre-processing—and a “gaze” function that mimics human eyesight.
The new system is based on RoboSense’s MEMS solid-state LIDAR, which merges 2D imagery with the company’s LIDAR algorithm developed specifically for autonomous driving—creating a high-resolution color point cloud.
To perform its gaze function, the RS-IPLS algorithm adjusts the Region of Interest (ROI) detection area in the point cloud—in real time with no delay. The ROI is a subset of an image identified for a particular purpose: in this case, nearby obstacles the self-driving vehicle needs to navigate around.
The RS-ILPS software will identify obstacles, such as an oncoming cyclist or the open door of a parked car, and initiate a “gaze” processing mechanism that locks the target for ROI processing. This gives the self-driving vehicle a clear, stable and dynamic read on its surrounding environment. And like a human driver, it actively monitors those targets while performing routine driving actions—the way a human driver might pay more attention to the cyclist while still keeping within traffic lanes.
The RS-IPLS system also provides richer three-dimensional spatial data information in real time. Data pre-processing is performed by the AI algorithm, with the ROI repeatedly detected for farther detection distance and more accurate perception results—with reduced stress on the software for data processing.
LiDAR technology is a key factor in the price and reliability of autonomous vehicles. Traditional LiDAR sensors are expensive to build, hard to miniaturize and can be sensitive to a moving car’s bumps and vibrations. Solid-state LiDAR is a promising solution to this problem, being more stable, safer and cheaper to manufacture—while mechanical LiDAR products can cost as much as $80,000, the RS-IPLS is priced at only $200!
“The new RoboSense RS-IPLS […] is intelligent, accurate, lower cost, and faster—and is poised to take over the market,” said RoboSense’s co-founder Mark Qiu.
If the company’s solid-state LiDAR can perform as advertised, it could mean that affordable self-driving vehicles are that much closer to appearing in local car dealerships. And with its gaze function actively tracking potential dangers, it could make autonomous cars safer for passengers and pedestrians alike.
How LiDAR is used by autonomous vehicles.
Read more about breakthroughs in LIDAR technologies for autonomous driving at First Solid-State LiDAR Unmanned Vehicle Is Unveiled in China.