Detection and distance measurement using sensors is not always accurate. Sensor fusion makes up for this shortcoming by reducing inaccuracies. This study, therefore, proposes an extended Kalman filter (EKF) that reflects the distance characteristics of lidar and radar sensors. The sensor characteristics of the lidar and radar over distance were analyzed, and a reliability function was designed to extend the Kalman filter to reflect distance characteristics. The accuracy of position estimation was improved by identifying the sensor errors according to distance. Experiments were conducted using real vehicles, and a comparative experiment was done combining sensor fusion using a fuzzy, adaptive measure noise and Kalman filter. Experimental results showed that the study’s method produced accurate distance estimations.

Object detection is one of the vital components used for autonomous navigation in dynamic environments. Camera and lidar sensors have been widely used for efficient object detection by mobile robots. However, they suffer from adverse weather conditions in operating environments such as sun, fog, snow, and extreme illumination changes from day to night. The sensor fusion of camera and lidar data helps to enhance the overall performance of an object detection network. However, the diverse distribution of training data makes the efficient learning of the network a challenging task. To address this challenge, we systematically study the existing visual and lidar features based on object detection methods and propose an adaptive feature attention module (AFAM) for robust multisensory data fusion-based object detection in outdoor dynamic environments. Given the camera and lidar features extracted from the intermediate layers of EfficientNet backbones, the AFAM computes the uncertainty among the two modalities and adaptively refines visual and lidar features via attention along the channel and the spatial axis. The AFAM integrated with the EfficientDet performs the adaptive recalibration and fusion of visual lidar features by filtering noise and extracting discriminative features for an object detection network under specific environmental conditions. We evaluate the AFAM on a benchmark dataset exhibiting weather and light variations. The experimental results demonstrate that the AFAM significantly enhances the overall detection accuracy of an object detection network.

Object detection is an important factor in the autonomous driving industry. Object detection for autonomous vehicles requires robust results, because various situations and environments must be considered. A sensor fusion method is used to implement robust object detection. A sensor fusion method using a network should effectively meld two features, otherwise, there is concern that the performance is substantially degraded. To effectively use sensors in autonomous vehicles, data analysis is required. We investigated papers in which the camera and LiDAR data change for effective fusion. We propose a feature switch layer for a sensor fusion network for object detection in cameras and LiDAR. Object detection performance was improved by designing a feature switch layer that can consider its environment during network feature fusion. The feature switch layer extracts and fuses features while considering the environment in which the sensor data changes less than during the learning network. We conducted an evaluation experiment using the Dense Dataset and confirmed that the proposed method improves the object detection performance.

In traffic surveillance systems, accurate camera–LiDAR calibration is critical for effective detection and robust environmental recognition. Due to the significant distances at which sensors are positioned to cover extensive areas and minimize blind spots, the calibration search space expands, increasing the complexity of the optimization process. This study proposes a novel target-less calibration method that leverages dynamic objects, specifically, moving vehicles, to constrain the calibration search range and enhance accuracy. To address the challenges of the expanded search space, we employ a genetic algorithm-based optimization technique, which reduces the risk of converging to local optima. Experimental results on both the TUM public dataset and our proprietary dataset indicate that the proposed method achieves high calibration accuracy, which is particularly suitable for traffic surveillance applications requiring wide-area calibration. This approach holds promise for enhancing sensor fusion accuracy in complex surveillance environments.

Cameras and LiDAR sensors have been used in sensor fusion for robust object detection in autonomous driving. Object detection networks for autonomous driving are often trained again by adding or changing datasets aimed at robust performance. Repeat training is necessary to develop an efficient network module. Existing efficient network module development changes to hand design and requires much module design experience. For this, a neural architecture search was designed, but it takes much time and requires optimizing the design process. To solve this problem, we propose a two-stage optimization method for the offspring generation process in a neural architecture search based on reinforcement learning. In addition, we propose utilizing two split datasets to solve the fast convergence problem as the objective function of the genetic algorithm: source data (daytime, sunny) and target data (day/night, adversary weather). The proposed method is an efficient module generation method requiring less time than the NSGA-NET. We confirmed the performance improvement and the convergence speed reduction using the Dense dataset. Through experiments, it was proven that the proposed method generated an efficient module.

Traffic surveillance systems are essential for ensuring public safety and optimizing urban traffic flow by accurately detecting, classifying, and monitoring traffic law violations such as illegal parking and jaywalking. However, detecting such violations in dynamic urban environments poses significant challenges due to unpredictable conditions. Traditional camera-based systems excel in capturing visual details but suffer performance degradation under low light, occlusion, or complex intersection scenarios, leading to false positives and missed detections. Therefore, we propose a camera-LiDAR fusion-based traffic violation detection system. LiDAR's independence from lighting conditions complements camera capabilities, especially in night scenarios or occluded environments. Experimental results in a real-world urban environment demonstrate the system’s effectiveness in reducing false positives and missed detections compared to traditional camera-only approaches. This study underscores the potential of sensor fusion to improve traffic management systems and points out future research opportunities in challenging conditions.