As self-driving cars have been developed targeting level 4 and 5 autonomous driving, the capability of the vehicle to handle environmental effects has been considered importantly. The sensors installed on autonomous vehicles can be easily affected by blockages (e.g., rain, snow, dust, fog, and others) covering the surface of them. In a virtual environment, we can safely observe the behavior of the vehicle and the degradation of the sensors by blockages. In this letter, the CARLA simulator-based evaluation framework has been developed and the assessment of lane detection performance under sensor blockage by heavy rain, which was analyzed by using the experimental data. Thus, we thoroughly note that the accuracy of lane detection for the autonomous vehicle has been decreased as the rainfall rate increases, and the impact of the blockage is more critical to curved lanes than straight lanes. Finally, we have suggested a critical rainfall rate causing safety failures of the autonomous vehicles, based on reasonably established rainfall equation based on experimental rain datasets.

Autonomous driving includes recognition, judgment, and control technologies, and is implemented using sensors such as cameras, LiDAR, and radar. However, recognition sensors are exposed to the outside environment and their performance may deteriorate because of the presence of substances that interfere with vision, such as dust, bird droppings, and insects, during operation. Research on sensor cleaning technology to solve this performance degradation has been limited. This study used various types and concentrations of blockage and dryness to demonstrate approaches to the evaluation of cleaning rates for selected conditions that afford satisfactory results. To determine the effectiveness of washing, the study used the following criteria: washer, 0.5 bar/s and air, 2 bar/s, with 3.5 g being used three times to test the LiDAR window. The study found that blockage, concentration, and dryness are the most important factors, and in that order. Additionally, the study compared new forms of blockage, such as those caused by dust, bird droppings, and insects, with standard dust that was used as a control to evaluate the performance of the new blockage types. The results of this study can be used to conduct various sensor cleaning tests and ensure their reliability and economic feasibility.

Lane segmentation and Lane Keeping Assist System (LKAS) play a vital role in autonomous driving. While deep learning technology has significantly improved the accuracy of lane segmentation, real-world driving scenarios present various challenges. In particular, heavy rainfall not only obscures the road with sheets of rain and fog but also creates water droplets on the windshield or lens of the camera that affects the lane segmentation performance. There may even be a false positive problem in which the algorithm incorrectly recognizes a raindrop as a road lane. Collecting heavy rain data is challenging in real-world settings, and manual annotation of such data is expensive. In this research, we propose a realistic raindrop conversion process that employs a contrastive learning-based Generative Adversarial Network (GAN) model to transform raindrops randomly generated using Python libraries. In addition, we utilize the attention mask of the lane segmentation model to guide the placement of raindrops in training images from the translation target domain (real Rainy-Images). By training the ENet-SAD model using the realistically Translated-Raindrop images and lane ground truth automatically extracted from the CARLA Simulator, we observe an improvement in lane segmentation accuracy in Rainy-Images. This method enables training and testing of the perception model while adjusting the number, size, shape, and direction of raindrops, thereby contributing to future research on autonomous driving in adverse weather conditions.

This study is the first to develop technology to evaluate the object recognition performance of camera sensors, which are increasingly important in autonomous vehicles owing to their relatively low price, and to verify the efficiency of camera recognition algorithms in obstruction situations. To this end, the concentration and color of the blockage and the type and color of the object were set as major factors, with their effects on camera recognition performance analyzed using a camera simulator based on a virtual test drive toolkit. The results show that the blockage concentration has the largest impact on object recognition, followed in order by the object type, blockage color, and object color. As for the blockage color, black exhibited better recognition performance than gray and yellow. In addition, changes in the blockage color affected the recognition of object types, resulting in different responses to each object. Through this study, we propose a blockage-based camera recognition performance evaluation method using simulation, and we establish an algorithm evaluation environment for various manufacturers through an interface with an actual camera. By suggesting the necessity and timing of future camera lens cleaning, we provide manufacturers with technical measures to improve the cleaning timing and camera safety.

Pollutants degrade the performance of LiDAR sensors used in autonomous vehicles. Therefore, there is an urgent need to develop cleaning technology for these sensors. In this study, a solid-state LiDAR sensor was selected as a target and sprayed/dried with 2.5 g of a mixture of Arizona dust and Kaolin. To achieve optimal LiDAR cleaning performance, the washer pressure, spray time, spray angle, and target point were selected as major variables. Additionally, an optimal cleaning solution for each spray was formed via the design of experiments and optimization techniques. Model suitability was observed for the second spray through to the fourth. The cleaning rate increased with the washer pressure and spray time. The influence of these variables decreased as the number of sprays increased. The spray angle and target point exhibited no significant influence, but excellent cleaning was observed in some central areas. Verification test results were within 3% for the second through fourth sprays, indicating reliability. This study used a designed experiment with 30 scenarios to reveal optimized conditions for protecting the sensor performance from external visibility obstructions. Disclosing the optimization method lowers the barrier for sensor cleaning manufacturers to develop their own technology, which ultimately enhances safer and more efficient autonomous driving.

Light detection and ranging (LiDAR) provides three-dimensional environmental information that is critical for maintaining the safety and reliability of autonomous driving systems. However, dust accumulation on the LiDAR window can cause detection errors and degrade performance. This study determined the optimal spray conditions for accumulating dust to evaluate LiDAR sensor cleaning performance. A primary optimization experiment using spray pressure, spray speed, spray distance, and the number of sprays as variables showed that spray pressure and number of sprays had the most significant influence on the kinetic energy and distribution of dust particles. Notably, the interaction between spray distance and number of sprays—related to curvature effects—was identified as a key variable increasing process sensitivity. A supplementary experiment, which added spray angle as a variable, indicated that while spray pressure remained the most significant factor, spray angle and number of sprays had an indirect influence through interaction terms. Both experiments used the same response variable (point cloud data) interactions to stepwise analyze particle transfer and spatial diffusion. The resulting optimal conditions offer a standard basis for evaluating LiDAR cleaning performance and may help improve cleaning efficiency and maintenance strategies.