LiDAR Start-up Claims 1km Object Tracking Range

A LiDAR specialist is claiming its sensor technology is now able to recognize and track moving objects more than a kilometer away.

Silicon Valley start-up AEye, which develops LiDAR hardware, software and algorithmic solutions for autonomous vehicles (AV), says it has achieved a new range and scan rate record to improve the safety of self-driving vehicles. The company has established a series of tests to establish and verify the range and scan rates of its system dubbed iDAR.

To test range, a 20-foot long moving truck was tracked and continuously scanned down the length of a nearly 3,000-foot runway. At the end of the runway, the software-driven sensor system was, according to the company, able to continuously detect, and track the movements of the vehicle as well as detect runway signs and markers en route.

The company also claims iDAR helps enable perception to classify and track objects at a rate faster than the human eye, this includes the ability to calculate the vector and velocity of each object within a frame. During testing, the AEye sensor vehicle was also taken off the runway to extend the available test range to over 3,280 feet, where the system is said to have continued to track the truck without difficulty.

In a statement, AV research specialist VSI Labs senior engineer Sara Sargent called the results “impressive”, noting the truck was “clearly identifiable and visible” one kilometer downrange. “We were also able to verify that AEye’s iDAR system achieves scan rates of 100Hz. This is an outstanding achievement that demonstrates the true potential of perception systems to reliably and accurately detect and track objects at great range.”

The company also announced it raised a Series B funding of $40M from Taiwania Capital and existing investors including Kleiner Perkins, Intel Capital, Airbus Ventures and Tychee Partners, bringing total investment in the firm to more than $60M.

AEye claims its LiDAR technology can perform object classification and recognition at the sensor level within the same frame, bypassing hundreds of milliseconds of latency seen in currently deployed systems. The main advantages of LiDAR are its range resolution, angular discrimination and ability to classify objects, offset by its high cost and poor weather performance in heavy rain and snow.

Like many start-ups in the auto industry, AEye is using LiDAR but CEO and founder Luis Dussan, an aerospace and defense veteran, was adamant about not using lasers in the 900-nanometre wavelength, which poses a number of potential hazards. In October the company announced it had made 300m+ class solid-state LiDAR sensors scan a driver’s surroundings at rates of 100Hz or more. Scanning speeds currently achieved are typically around 10Hz.

The company claims it’s been able to do this as a result of the artificial intelligence system embedded in iDAR, which is designed to mimic, and outperform, the capabilities of the human visual cortex.

In a June interview with TU-Auto, AEye chief scientist Dr Allan Steinhardt said he believed that LiDAR combines the best of camera and radar capabilities, possessing both the naturalness of regular camera information and also the benefit of range. He explained that while early consumer solutions were expensive and ill matched to demanding automotive environmental requirements, the strength of innovation in the consumer auto space is driving improved solutions with lower costs.

— Nathan Eddy is a filmmaker and freelance journalist based in Berlin. Follow him on Twitter.


Leave a comment

Your email address will not be published. Required fields are marked *