New radar sensors can accurately see through smoke, dust, and fog

Researchers at Örebro University have now succeeded in increasing the accuracy of radar sensors for navigation to the point where the sensors may be utilized in autonomous automobiles, enabling safe driving in any condition, according to a study.

Autonomous vehicles use a variety of sensors and technologies to "see" their surroundings and navigate safely, including cameras, LIDAR (Light Detection and Ranging), radar (Radio Detection and Ranging) and ultrasonic sensors.

All these sensors provide data to the autonomous vehicle's computer system, which uses machine learning algorithms and other technologies to process the data and make decisions about navigating the environment. Combining the data from these various sensors allows autonomous vehicles to comprehensively view their surroundings and make real-time decisions to avoid obstacles, stay within lanes, and follow traffic laws.

“The advantage with radar is that it works in all weather conditions and can ‘see’ through smoke and dust,” says Daniel Adolfsson, a doctoral student in computer science at Örebro University.

With this new method, an autonomous vehicle can pinpoint its location with a precision of 1 meter (3.28 feet) after traveling 100 meters (328 feet). That enhances the positioning system of the radar sensors by 1 meter (3.28 feet).

“Reducing the error margin from 2 to 1 percent is a huge step forward. The method is both very quick and precise, which is just the ticket if autonomous robots are to interact safely with humans and other robots,” says Daniel Adolfsson.

“Our work with improving the precision of radar sensors can lead to autonomous cars being able to drive safely no matter the weather conditions. It can also prove useful within the construction and mining industries where autonomous heavy-duty machinery must be able to operate in environments with a lot of dust,” Adolfsson added.

Using this new method, laser location is becoming increasingly precise compared to radar positioning. Radar sensors can see through smoke, dust, and fog, so they can replace laser sensors on vehicles that operate in low visibility.

Autonomous vehicles seem to be getting safer and safer with each development and are more prevalent in the public eye.

Abstract:

This article presents an accurate, highly efficient, and learning-free method for large-scale odometry estimation using spinning radar, empirically found to generalize well across very diverse environments—outdoors, from urban to woodland, and indoors in warehouses and mines—without changing parameters. Our method integrates motion compensation within a sweep with one-to-many scan registration that minimizes distances between nearby oriented surface points and mitigates outliers with a robust loss function. Extending our previous approach, conservative filtering for efficient and accurate radar odometry (CFEAR), we present an in-depth investigation on a wider range of datasets, quantifying the importance of filtering, resolution, registration cost and loss functions, keyframe history, and motion compensation. We present a new solving strategy and configuration that overcomes previous issues with sparsity and bias, and improves our state-of-the-art by 38%, thus, surprisingly, outperforming radar simultaneous localization and mapping (SLAM) and approaching lidar SLAM. The most accurate configuration achieves 1.09% error at 5 Hz on the Oxford benchmark, and the fastest achieves 1.79% error at 160 Hz.