Association of Multi-sensor Data for Autonomous Car Driving: A Comparative Evaluation

Summary: Autonomous driving has the potential to revolutionize transportation by making it safer, more efficient, and more useful. However, calls on the automobile to be able to detect and understand its surroundings which may be challenging given the complexity and variety of real-world events. The integration of multisensory data plays a crucial role in enhancing the capabilities of autonomous vehicles to perceive their environment accurately and in making informed decisions. An autonomous car's environment can't be analyzed fully and accurately by a single sensor, so multisensory data integration is essential. It is easy for cameras to detect visual information like lane markings and traffic signals, but poor lighting or occlusions can affect their performance. However, lidar and radar sensors may struggle to detect certain objects such as pedestrians or animals due to their lack of accuracy and sensitivity to lighting conditions. To get beyond these constraints, multisensory data integration methods like sensor fusion have been investigated. These methods combine input from many sensors to produce a more accurate representation of the world. This can be achieved using a variety of techniques, such as probabilistic filtering, machine learning, and geometric fusion. For multisensory data integration, synchronization and calibration of the data are major obstacles. To build an accurate depiction of the environment, the data from several sensors must be time-synchronized. For the data to be merged accurately, it may also be necessary to alter the calibration settings of the sensors, such as their field of view or resolution. Multisensory data is extremely advantageous for autonomous vehicle driving. For instance, combining several sensors can increase object tracking's precision and dependability, which lowers the risk of mishaps. Additionally, by providing more details about the environment, such as traffic congestion, weather conditions, and road conditions, the utilization of multisensory data can improve the user experience. The incorporation of multisensory data can also increase the overall effectiveness of autonomous vehicle driving which results in lower pollutants and fuel usage. This report presents a comprehensive analysis of various papers related to the association of multisensory data for autonomous car driving. The analysis focuses on the methods, techniques, and algorithms employed to fuse data from different sensors such as cameras, LiDAR, radar, and GPS which helps to improve the vehicle's perception of its surroundings. The Key aspects evaluated include the effectiveness of sensor fusion in enhancing object detection, tracking and environment mapping as well as the impact on overall system performance, reliability, and safety.