Determination of Traffic Lane in Tunnel and Positioning of Autonomous Vehicles Using Chromaticity of LED Lights

Currently, the location recognition and positioning system are the essential parts of unmanned vehicles. Among them, location estimation under GPS-denied environments is currently being studied using IMU, Wi-Fi, and VLC, but there are problems such as cumulative errors, hardware complexity, and precision positioning. To address this problem with the current positioning system, the present study proposed a lane positioning technique by analyzing the chromaticity coordinates, judging from the color temperature of LED lights in tunnels. The tunnel environment was built using LEDs with three color temperatures, and to solve nonlinear problems such as lane positioning from chromaticity analysis, a single input single output fuzzy algorithm was developed to estimate the position of an object on lanes using chromaticity values of signals measured by RGB sensors. The RGB value measured by the sensor removes the disturbance through the pre-processing filter, accepts only the tunnel LED information, and estimates where it is located on the x-distance indicating the lane position through a fuzzy algorithm. Finally, the performance of the fuzzy algorithm was evaluated through experiments, and the accuracy was shown with an average error of less than 4.86%.

Automated summary

Location recognition and positioning systems are essential for unmanned vehicles. However, there are challenges in estimating location in GPS-denied environments, such as cumulative errors, hardware complexity, and precision positioning. This study proposes a lane positioning technique using chromaticity analysis of LED lights in tunnels and a fuzzy algorithm for accurate estimation.