Vehicle localization in highway tunnels is challenging for autonomous vehicle navigation. The drift error increases because GPS signals from satellites cannot be received inside a highway tunnel. Therefore, map-aided localization is essential for correcting the drift error. However, the environment around the tunnel is composed mainly of elliptical walls, and there are few unique feature points for map matching, unlike the case outdoors. In this paper, we propose a tunnel facility–based precise vehicle localization method for highway tunnels using 3D LIDAR. For vehicle localization, a point landmark map storing the center points of the tunnel facilities and a probability distribution map storing the probability distributions of the lane markings is used. In the proposed method, the tunnel facility points are extracted using 3D LIDAR, and position estimation is conducted using an extended Kalman filter-based navigation filter. The proposed localization algorithm was verified through experiments using actual highway driving data. The experimental results verify that tunnel facility–based vehicle localization yields precise results in various situations.

We propose a free-resolution probability distributions map (FRPDM) and an FRPDM-based precise vehicle localization method using 3D light detection and ranging (LIDAR). An FRPDM is generated by Gaussian mixture modeling, based on road markings and vertical structure point cloud. Unlike single resolution or multi-resolution probability distribution maps, in the case of the FRPDM, the resolution is not fixed and the object can be represented by various sizes of probability distributions. Thus, the shape of the object can be represented efficiently. Therefore, the map size is very small (61 KB/km) because the object is effectively represented by a small number of probability distributions. Based on the generated FRPDM, point-to-probability distribution scan matching and feature-point matching were performed to obtain the measurements, and the position and heading of the vehicle were derived using an extended Kalman filter-based navigation filter. The experimental area is the Gangnam area of Seoul, South Korea, which has many buildings around the road. The root mean square (RMS) position errors for the lateral and longitudinal directions were 0.057 m and 0.178 m, respectively, and the RMS heading error was 0.281°.

Road markings are always present on roads to guide and control traffic. Therefore, they can be used at any time for vehicle localization. Moreover, they can be easily extracted by using light detection and ranging (LIDAR) intensity because they are brightly colored. We propose a vehicle localization method using a 2D road marking grid map. The grid map inserts the map information into the grid directly. Thus, an additional process (such as line detection) is not required and there is no problem due to false detection. We obtained road marking using a 3D LIDAR (Velodyne HDL-32E) and binarized this information to store in the map. Thus, we could reduce the map size significantly. In the previous research, the road marking grid map was used only for position estimation. However, we propose a position-and-heading estimation algorithm using the binary road marking grid map. Accordingly, we derive more precise position estimation results. Moreover, position reliability is an important factor for vehicle localization. Autonomous vehicles may cause accidents if they cannot maintain their lane momentarily. Therefore, we propose an algorithm for evaluating map matching results. Consequently, we can use only reliable matching results and increase position reliability. The experiment was conducted in Gangnam, Seoul, where GPS error occurs largely. In the experimental results, the lateral root mean square (RMS) error was 0.05 m and longitudinal RMS error was 0.08 m. Further, we obtained a position error of less than 50 cm in both lateral and longitudinal directions with a 99% confidence level.

This paper describes a fully automated driving algorithm on complex urban roads with LiDAR, vision, and GPS/map based environment representation with guaranteed safety. The proposed algorithm consists of the following three steps: an environment representation, a safety driving envelope decision, and a motion optimization. An environment representation system consists of three main modules: object classification, vehicle/non-vehicle tracking and map/lane based localization. A motion planning modules derives an optimal motion as a function of time, from the environment representation results. A safety envelope decision module determines the complete driving corridor that leads to the destination while assigning all objects to either the left or right corridor bound. In the case of moving objects such as other traffic participants, their behaviors are anticipated in the near future. A motion optimization module uses the safety envelop as a constraint and computes a trajectory that the vehicle stays in its bounds. The vehicle control module feeds back the pose estimate of the localization module to guide the vehicle along the planned trajectory. The effectiveness of the proposed automated driving algorithm is evaluated via vehicle tests. Test results show the robust performance on an inner-city street scenario.

For the safe driving of autonomous vehicles, accurate position estimation is required. Generally, position error must be less than 1m because of lane keeping. However, GPS positioning error is more than 1m. Therefore, we must correct this error and a map matching algorithm is generally used. Especially, road marking intensity map have been used in many studies. In previous work, 3D LIDAR with many vertical layers was used to generate a local intensity map. Because it can be obtained sufficient longitudinal information for map matching. However, it is expensive and sufficient road marking information cannot be obtained in rush hour situations. In this paper, we propose a localization algorithm using an accumulated intensity local map. An accumulated intensity local map can be generated with sufficient longitudinal information using 3D LIDAR with a few vertical layers. Using this algorithm, we can also obtain sufficient intensity information in rush hour situations. Thus, it is possible to increase the reliability of the map matching and get accurate position estimation result. In the experimental result, the lateral RMS position error is about 0.12m and the longitudinal RMS error is about 0.19m.