Centralised and Decentralised Sensor Fusion-Based Emergency Brake Assist

Abstract

Many advanced driver assistance systems (ADAS) are currently trying to utilise multi-sensor architectures, where the driver assistance algorithm receives data from a multitude of sensors. As mono-sensor systems cannot provide reliable and consistent readings under all circumstances because of errors and other limitations, fusing data from multiple sensors ensures that the environmental parameters are perceived correctly and reliably for most scenarios, thereby substantially improving the reliability of the multi-sensor-based automotive systems. This paper first highlights the significance of efficiently fusing data from multiple sensors in ADAS features. An emergency brake assist (EBA) system is showcased using multiple sensors, namely, a light detection and ranging (LiDAR) sensor and camera. The architectures of the proposed ‘centralised’ and ‘decentralised’ sensor fusion approaches for EBA are discussed along with their constituents, i.e., the detection algorithms, the fusion algorithm, and the tracking algorithm. The centralised and decentralised architectures are built and analytically compared, and the performance of these two fusion architectures for EBA are evaluated in terms of speed of execution, accuracy, and computational cost. While both fusion methods are seen to drive the EBA application at an acceptable frame rate (~20fps or higher) on an Intel i5-based Ubuntu system, it was concluded through the experiments and analytical comparisons that the decentralised fusion-driven EBA leads to higher accuracy; however, it has the downside of a higher computational cost. The centralised fusion-driven EBA yields comparatively less accurate results, but with the benefits of a higher frame rate and lesser computational cost.

Small Object Detection and Tracking: A Comprehensive Review

“…Sensor fusion is a popular approach in modern perception systems, where information from multiple sensors is combined to enhance the overall perception and understanding of the environment. In the context of object detection and tracking, integrating 3D LiDAR data with other sensors, such as cameras, can provide more accurate and comprehensive information about the surrounding objects [ 90 , 91 ].…”

Switching Trackers for Effective Sensor Fusion in Advanced Driver Assistance Systems

“…In this paper, for the purpose of gauging the performance of the tracker algorithm under multiple scenarios, we have used some standard off-the-shelf LiDAR, camera, and RaDAR detection algorithms, namely, YOLOv4 for camera object detection as implemented by Kumar et al [9], a RaDAR based object detection methodology as worked upon by Manjunath et al [10], and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) for LiDAR object detection, which is similar to that developed by Deng et al [11]. As explained earlier in Section 1, we shall use the target-level sensor fusion model, as worked upon by Deo et al [5], for fusing the sensor data.…”

“…Several studies on the topic of multi-object sensor fusion and tracking are available in the public domain [2,4,5]. In a typical target-level sensor fusion architecture, all target objects are independently tracked once the data from multiple sensors is associated or fused [6].…”

“…Sensor fusion techniques, on the other hand, leverage the output provided by multiple sensors; as a result, they are more reliable and error-free, compared to the techniques that employ a single sensor to gauge the environment [2,3]. Several studies on the topic of multi-object sensor fusion and tracking are available in the public domain [2,4,5]. In a typical target-level sensor fusion architecture, all target objects are independently tracked once the data from multiple sensors is associated or fused [6].…”

“…Considering this problem statement, in this paper, we build a logic which switches between the Extended Kalman Filter and the Unscented Kalman Filter, based on the realworld traffic density surrounding the ego-vehicle, and the ego-vehicle's type of Overall, there are a variety of algorithms, present in today's technological world, that may be used for object detection, fusion, and tracking. While there are multiple fusion architectures, for our work, we follow a target-level sensor fusion architecture (also known as centralised fusion architecture) for ease of application and experimentation, as described in the work presented by Deo et al [5]. There are multiple works which provide data regarding which tracking algorithms work best under specific external scenarios.…”