New Stealth Camera Tech Replaces LiDAR to Tackle Laser Detection

image provided by pixabay

This post is also available in: עברית (Hebrew)

Researchers from the Southwest Research Institute in Texas developed camera-based autonomous driving tools that can work without using technologies like LIDAR and RADAR, an innovation that can be useful to military stealth capabilities, as well as in applications for space and agriculture.

Modern autonomous driving solutions use light detection and ranging sensors (LiDAR) to visualize objects around the vehicle. High-energy lasers are sent out in all directions to determine the depth and distance from various objects. The data is then sent to a software solution that identifies the objects and helps the vehicle’s computer decide if it should stop or slow down.

While this technology works for civilian transport, the lights from the LiDAR sensors are easy to detect, which could easily compromise the location of a unit if used on military vehicles. This is similar to radar technology that uses radio waves (which are easy to detect) and GPS that uses satellites (which can be blocked easily).

According to Interesting Engineering, the researchers at SwRI responded to this problem by creating Vision for Off-Road Autonomy (VORA), a suite of tools to perceive objects passively, create environment models, and localize units on a map, even in off-road environments.

The researchers created this system as an alternative to LiDAR that uses cameras since they do not emit light or laser signals. However, to use the camera data to perform the needed high-precision tasks, the team had to develop new software. And so, they created a deep learning stereo matcher (DLSM), a neural network-based tool to create dense maps using the camera data by using disparity or differences between the motion in the stereo images to create highly accurate maps.

When it comes to applications of this innovation, the project was originally meant for the defense clients of the Southwest Research Institute. However, the researchers saw that VORA’s capabilities could be used for many other applications. It could, for example, be deployed in planetary exploration, where autonomous vehicles on other planets could benefit from the energy efficient camera-based system. It could also be used to aid agriculture automation in regions where GPS signals are hard to reach, like canyons and mountainous slopes.