Sensing Structure for Blind Spot Detection System in Vehicles

In today's world of automotive, one of the key innovative features and technology advancement that we could see on automobiles or cars is by having an external or internal camera or automotive camera. The said cameras can be installed on either or both the front and rear sides of a vehicle for safety purposes. To name a few on the basis of the automotive camera application are park assist system, advanced driver assistant monitoring system (ADAS), blind spot detection and night vision system. Because of the safety concerns, all produced cameras must be free from any particle defect especially not blocking its image seen by the user. Particle defect is one of the major issues seen on cameras as it entails lots of problems for the user. Thus, this study aims to reduce and identify a particle defect during the camera production which prevent it to send to the OEM (Overall Equipment Manufacturer) or the end user. The particle defect data collected is based on the daily, weekly and monthly production of the automotive camera's scrap monitoring. DMAIC (Define-Measure-Analyze-Improve-Control) approach is used as a structured methodology to understand the particle defect and reducing it by improving the process. This approach also provides decisions that is based on the statistical analysis using Two Proportions test (or Z-test). There are seven validated significant factors that are true causes namely particle escape from vacuum filter, one lens cleaning chamber used for different processes, damaged blister trays from supplier, current alignment tester area is open for man and material movements, upright position of the module during lens cleaning, airborne particle dropped imager during staging and customer specification is too tight for particle criteria. The result of getting the major root cause of this particle defect leads to the researcher to offer and recommend strategies on how to prevent camera particle scrap reduction and improvement process of the automotive camera.

Blind spot detection is an important feature of Advanced Driver Assistance Systems (ADAS). In this paper, we provide a camera-based deep learning method that accurately detects other vehicles in the blind spot, replacing the traditional higher cost solution using radars. The recent breakthrough of deep learning algorithms shows extraordinary performance when applied to many computer vision tasks. Many new convolutional neural network (CNN) structures have been proposed and most of the networks are very deep in order to achieve the state-of-art performance when evaluated with benchmarks. However, blind spot detection, as a real-time embedded system application, requires high speed processing and low computational complexity. Hereby, we propose a novel method that transfers blind spot detection to an image classification task. Subsequently, a series of experiments are conducted to design an efficient neural network by comparing some of the latest deep learning models. Furthermore, we create a dataset with more than 10,000 labeled images using the blind spot view camera mounted on a test vehicle. Finally, we train the proposed deep learning model and evaluate its performance on the dataset.

In recent years, the country has gradually paid attention to the intelligent and connected automobile industry, from emergency driving to the current auxiliary driving, the blind spot detection system is also gradually upgraded. For the early warning BSD system, how to have a better alarm signal definition and a higher accuracy is the basis of upgrade direction. In the actual driving process, the change in road conditions and the driver’s negligence make the BSD system has the upgrade space. First of all, according to the BSD test of GB / T 39265-2020, evaluates its rationality and propose an improvement plan. The evaluation scheme of key functions that can be applied to blind spot monitoring in L2 driving assistance is designed to effectively guarantee the further standardization and upgrading of the blind spot detection system.

Car accidents often happen when vehicles change lanes, turn at intersections, back up, or park in a parking lotthat is especially dangerous. Therefore, detecting people in the blind spots of cars is a significant issue. There are many methods to detect people. However, those problems do not completelysolve the problem. To solve the problem, the article proposes a system to recognize people entering the blind spot behind acar. The method that we use is to combine image processing algorithms with microwave radar sensors to collect information about the vehicle's surroundings. We use the Yolo-fastest module to extract human detection through the parts of the human body (arms, legs, head, torso), combined to increase recognition ability. Besides, using microwave radar sensors are used to detect moving objects. The results on cars have achieved 91.63% accuracy and have potential applications in self-driving cars and advanced driver assistance systems.

Deep learning-based driver assistance systems (ADAS) have attracted interest from researchers due to their impact on improving vehicle safety and reducing road traffic accidents. In Uganda, road accidents have continued to soar with an increase of up to 42% in 2021 due to the growing road traffic density. To curb the high rates of road accidents, especially for heavy-duty vehicles, Kiira Motors Corporation a state-owned mobility solutions enterprise needs advanced driver assistance systems for improved safety of their market entry products- the Kayoola buses. This research presents an approach to vehicular safety enhancement through the integration of Lane Departure Warning (LDW) and Blind Spot Detection systems (BSD) using advanced deep learning algorithms. The resultant LDW and BSD system is realized on the Raspberry Pi platform, incorporating diverse sensors. By combining these advanced features, the study not only bridges an essential research void but also offers a practical resolution to pressing road safety concerns in the East African context. The integration of LDW and BSD systems through deep learning techniques marks a pivotal advancement in vehicular safety. The lane detection model was tested on DET and TuSimple datasets. Our model attained a mean F1 Score of 77.59% and a mean IoU of 65.26% on the Dataset for Lane Extraction (DET) and an overall accuracy of 97.96% on the TuSimple dataset. Our work presents an integrated lane departure warning and blind spot detection system that will be able to alert the driver using the graphical user interface, and auditory feedback. The anticipated real-world implementation is poised to substantiate the system’s effectiveness, thereby contributing to safer roads regionally and inspiring innovation in automotive engineering by leveraging artificial intelligence.

Many tasks that until recently could only be performed by humans are now entrusted to machines. Automation plays an increasingly important role in many areas of life, and its special role is observed in broadly understood transport, affecting the mobility, behaviour, and safety of drivers. Faced with the threat of a large number of accidents, deaths, and injuries, vehicles equipped with driver-assistance systems, the so-called ADAS (advanced driver-assistance system), are a significant support for the driver in reacting to the unexpected situations. The studies conducted so far indicate that the experience of using these systems may increase not only the comfort but most of all the safety of road users. In Poland, such analyses are still a challenge. The aim of the study designed by scientists from the Motor Transport Institute was to analyse the impact of the experience in using a vehicle equipped with systems on the speedy adaptation to the use of new driver-assistance systems. The adaptation process for the use of three systems was taken into account: adaptive cruise control, blind spot detection system, and lane keeping assistant. The research was carried out in real traffic. Particular attention was paid to the reactions of people who have no experience with the use of driver-assistance systems in vehicles, compared to those who have used these systems before. Proprietary questionnaires on the drivers’ feelings were used for the analyses. As expected, the analyses showed that experience with driver-assistance systems is an important factor influencing the speed of adaptation and the sense of comfort in using the systems.

This paper details the design and implementation of an embedded real-time system to fuse data from two 77 GHz radars, three 24 GHz radars and a video lane detection system fitted to a vehicle to track the movement of other vehicles in the local environment. The system demonstrates excellent tracking of vehicles as they pass through regions covered by single and multiple sensors as well as regions with no sensor coverage where tracking continues despite no new measurements. Such a system facilitates future driver assistance functions such as lane change support and blind spot detection.

Objectives: In India, the frequency of traffic collisions is among the highest in the world. Most of the accidents are due to blind spots which cannot be directly observed by the driver at control. The paper proposes a collision avoidance system due to blind spots and in the case of heavy traffic scenarios. Methods: The proposed system avoids collision in blind spot and in heavy traffic and does not follow the path of preceding vehicle if the space between detector (Our Vehicle) and host vehicle is in proximity. For implementation of the system, Arduino boards and other drivers along with ultrasonic sensor are used and ZigBee is used for data communication between two nodes. Findings: The Blind Spot Detection (BSD) procedure based on wireless detecting technique ZigBee is proposed to monitor the blind spot region for the presence of interferences, automobiles, or other objects with the aid of ultrasonic sensor. For detecting the vehicles in blind spot, Neural Fuzzy Logic is used for the proposed system. The connection models for the developed system is identified and proposed. The paper also discusses about the advantages of ZigBee over other wireless technologies. The Wireless Technology identified is ZigBee which will lower the wiring connection in the vehicles. Applications: This method is useful to avoid collision in blind spot and also lowered the wiring connection in the vehicles by using ZigBee. Hence, the proposed system discovers its application in real time and gives collision free environment. The system is evaluated for different driving scenarios and results are discussed. Keywords: Blind Spot Detection, Collision-Avoidance System, ZigBee Pair

In this paper, we developed an integrated mid-range automotive radar system with both blind spot detection (BSD) and lane change assistance (LCA) functions. Integrated automotive radar extends the maximum detection range to 80m with an update time that is within 50ms. By means of ROI preprocessing technique, the computation load of radar signal processing was improved. In order to evaluate the detection performance considering an actual driving environment, radar performance evaluations were conducted in various driving scenarios at an automotive proving ground.

Over the last decade, new applications have emerged for radar technology in the automotive industry, such as adaptive cruise control, blind spot detection and automatic emergency braking. It is expected that the continued development of radar will unlock new capabilities for autonomous vehicles and safety systems. Indeed, advancements in integrated circuit design have enabled the development of very high-frequency radars with sophisticated signal processing and machine learning techniques. As a consequence, it is now possible with low power, small form factor and low-cost sensors to have rich point cloud data in dense environments. However, many challenges remain that can reduce precision, recall and accuracy of automotive radar, intra- and inter-platform interference or jamming, multi-radar fusion, multipath effects, false targets and detection ambiguity (range, velocity and angle). Many approaches have been developed to address these challenges; massive hybrid MIMO (Multiple-Input Multiple-Output), compressed sensing methods, sparse arrays, PMCW (Phase-Modulated Continuous Wave) Radar, Artificial Intelligence (AI) algorithms and others. Some of these approaches have already been applied to automotive radars, and some likely will be used in the coming years. This chapter will review why radar is unique, the challenges and the state-of-the-art solutions. We will use intuitive and simplified examples to provide the background needed to understand the utility of radar and these techniques.

Radar is a key technology in the safety system of a modern vehicle. Automotive radars are the critical sensors in advanced driver-assistance systems, which are used in adaptive cruise control, collision avoidance, blind spot detection, lane change assistance, and parking assistance. The book covers all the modern radars used in automotive technology. A long-range radar mounted in the front of the vehicle is usually for adaptive cruise control. The medium range radars mounted in the front and rear provide wider coverage than the long-range radars and they can be used for cross traffic alert and lane change assistance. The corner mounted short range radars support parking aid, obstacle/pedestrian detection and blind spot monitoring. In real applications, these radars usually work together to provide more robust detection results. In this book, we also recognize that the future of automotive radars should not only address conventional exterior applications, but also play important roles for interior applications, such as gesture sensing for human-vehicle interaction and driver/passenger vital signs and presence monitoring. The book is aimed at those radar engineers who are working on automotive applications.

This paper aims to thoroughly examine and compare advanced driver-assistance systems (ADASs) in the context of their impact on safety and driving comfort. It also sought to determine the level of acceptance and trust drivers have in these systems. The first chapter of this document describes the sensory detectors used in ADASs, including radars, cameras, LiDAR, and ultrasonics. The subsequent chapter presents the most popular driver assistance systems, including adaptive cruise control (ACC), blind spot detection (BSD), lane keeping systems (LDW/LKS), intelligent headlamp control (IHC), and emergency brake assist (EBA). A key element of this work is the evaluation of the effectiveness of these systems in terms of safety and driving comfort, employing a survey conducted among drivers. Data analysis illustrates how these systems are perceived and identified areas requiring improvements. Overall, the paper shows drivers’ positive reception of ADASs, with most respondents confirming that these technologies increase their sense of safety and driving comfort. These systems prove to be particularly helpful in avoiding accidents and hazardous situations. However, there is a need for their further development, especially in terms of increasing their precision, reducing false alarms, and improving the user interface. ADASs significantly contribute to enhancing safety and driving comfort. Yet, they are still in development and require continuous optimization and driver education to fully harness their potential. Technological advancements are expected to make these systems even more effective and user-friendly.

This paper proposes an overview of achievements in the safety benefits (lives extended and injuries mitigated) provided by current or close-to-market vehicle technologies. It also proposes to review some of the research and developments in future vehicle safety such as connected vehicles and automated driving. It finally discusses expected safety benefits, and potentials as well as barriers to a large dissemination of these current and future technology-based traffic safety measures. As they are mostly foreseen in high income countries in the middle term, another type of revolution is needed in low and middle income countries for before these technologies can be put onto all global markets at a reasonable cost and contribute to the safety benefits of all road users. Frugal safety, coming from the frugal engineering concept, would be the recommended way to enable and accelerate this deployment. As for vehicle safety technologies already on the market, they have proven to be highly effective in mitigating vehicle occupant injuries, and injuries of vulnerable road users. Secondary safety devices and basic driver assistance systems have been largely evaluated and the effectiveness of seat belts, load limiters, pretensioners, airbags, car structure, electronic stability control (ESC), emergency brake assist (EBA) and other such devices have been examined. Available results show that, for example, if all cars were Euro NCAP five stars and fitted with EBA and ESC, compared to four stars without ESC and EBA, injury accidents would be reduced by 47%, all injuries would be mitigated by 68% and severe + fatal injuries by 70% (Page et al. in 53th AAAM conference, Baltimore, 2009). Unlike secondary safety systems, ESC, ultrasonic park assist, cornering lights and manual speed limiters, most of the primary safety systems or advanced assistance systems (such as blind spot detection, lane departure warning, lane keeping assist, automated cruise control, night vision and high speed AEB) are not, or just poorly, deployed on the market. There are a few reasons for this. First, their maturity and their capacity to work well without too many counter effects are recent, technological barriers have been strong for a long time. They were launched first in luxury cars in the USA, Japan and Western Europe in the mid-2000s and democratization has started just a few years back. Secondly, these features are costly, in terms of unit price and also in terms of investment; and their deployment demands progress in cost reduction. Thirdly, their expected safety benefits are, to a certain extent, not really known. If it is now epidemiologically demonstrated that secondary safety features bring safety, as well as ESC, the positive effectiveness of not yet largely deployed primary safety devices is only estimated and not yet proven, except for some of them like low speed automatic emergency braking (AEB). Studies about effectiveness of these devices take into account the potential in saving lives of generic functions, but a lot is still unknown about the real-world usage and acceptance by drivers and pedestrians of these applications. As there are many variants of these features, and especially concerning the mode of restitution to the driver of the information, the alert, or the warnings, the genuine efficiency of each variant is still a mystery. Of course, in-house studies by suppliers and car manufacturers anticipate positive acceptance by drivers and have identified and countered possible counter effects. Of course, current research is evaluating their potential to safety and in covering real-world driver needs, but a large field of research is still open for their safety assessment in real-life. Anyhow, the introduction of vehicle safety technologies is supposed to bring safety for a variety of road users and a variety of crash configurations. Of course, they will first be disseminated in high income countries but they will also be further disseminated, expectedly quickly, in the rest of the world under the pressure of customer demand, competition between vehicle makers as well as pressure from all kinds of New Car Assessment Programs, NCAP’s (Latin NCAP in South America, Bharat NCAP in India, Asian NCAP in Asia, etc.), and more generally from Global NCAP, which federates NCAP’s all over the world. This will be a first step in the globalization of existing or close-to-development vehicle safety technologies. Now, further on, emerging (more complex) technologies such as connectivity (vehicle-to-X communications) and automated driving will bring additional opportunities to tackle road crashes by adding safety applications not yet covered by passive safety or traditional (advanced) assistance systems. Actually connected technologies for safety and driving assistance systems in the current world of stand-alone technologies have three additional values:

Smart cars or Autonomous Vehicles (AV), capable of replicating human cognitive behaviour supported with a digital autopilot, shall make better drivers than humans in the near future. Driving in a complex and competitive environment is a challenging task. Driver health monitoring serves to be a pivotal contributor to autonomous driving. The human factors, including cognitive, psychomotor and physiological abilities, should be continuously monitored in an autonomous vehicle. On the road, neurocognitive impairments are very often reported, where the driver suffers a stroke or heart attack or unconsciousness and is incapacitated from driving. In this paper, we present AI-NAAV (AI enabled Neurocognition Aware Autonomous Vehicle), a deep learning assisted AI-empowered autonomous vehicle, which has the combined capabilities of a) detecting impaired driving capabilities in the driver b) shift to autopilot mode and hence enabling a safe drive. AI-NAAV is simulated on CARLA platform. The exteroceptive sensor combinations used for surround view, blind spot detection, rear collision warning and parking assistance, are used in a complementary mode for a safe drive on autopilot. The paper discusses the sensors deployed, methodologies and the results from simulations that perform an accurate transition of autonomy from Level 0 to Level 5, as defined in SAE (Society of Automotive Engineers).

Advanced rider assistance systems (ARAS) play a crucial role in enhancing motorcycle safety through features such as collision avoidance, blind-spot detection, and adaptive cruise control, which rely heavily on sensors like radar, cameras, and LiDAR. However, their performance is often compromised under adverse weather conditions, leading to sensor interference, reduced visibility, and inconsistent reliability. This study evaluates the effectiveness and limitations of ARAS technologies in rain, fog, and snow, focusing on how sensor performance, algorithms, techniques, and dataset suitability influence system reliability. A thematic analysis was conducted, selecting studies focused on ARAS in adverse weather conditions based on specific selection criteria. The analysis shows that while ARAS offers substantial safety benefits, its accuracy declines in challenging environments. Existing datasets, algorithms, and techniques were reviewed to identify the most effective options for ARAS applications. However, more comprehensive weather-resilient datasets and adaptive multi-sensor fusion approaches are still needed. Advancing in these areas will be critical to improving the robustness of ARAS and ensuring safer riding experiences across diverse environmental conditions.