BSD - Blind Spot Detection System

According to a report from the Insurance Institute for Highway Safety, which analyzed data from the U.S. Department of Transportation's Fatality Analysis Reporting System, a total of 4,764 people lost their lives in crashes involving large trucks in 2022. Of these fatalities, 17% were truck occupants, 66% were occupants of cars and other passenger vehicles, and 16% were pedestrians, bicyclists, or motorcyclists. Given these statistics, addressing the issue of large truck-related deaths is crucial for improving road safety and protecting the well-being of all road users.

Figure 1: Data analysis of fatalities in large truck accidents

There are many factors that contribute to large truck traffic accidents, with one of the most significant being the truck driver's blind spot. Large trucks typically have larger blind spots, which can pose serious safety risks. As shown in the following figure:

Figure 2: Schematic diagram of the blind spot of a large truck, with the cabin on the left

This is particularly dangerous when turning. For example, trucks with the cabin on the right side often have limited visibility of the area to the left when making a left turn. Similarly, trucks with the cabin on the left side may struggle to fully assess the situation on the right side when turning right.

To address this issue and enhance road safety, Ultravision has developed an artificial intelligence-based blind spot detection system. This innovative technology aims to improve the safety of large vehicles, offering greater protection for drivers, passengers, and pedestrians, and contributing to the creation of a safer, more prosperous urban environment.

Technical Features

Ultravision's blind spot detection(BSD) system combines machine vision and artificial intelligence technology to enhance road safety. The system uses cameras to capture images of the blind spot and employs visual ranging to assess the risk within that area. By integrating human detection, object recognition, and other advanced technologies, the system can identify vulnerable road users (VRUs) —such as pedestrians, scooter riders, cyclists, and motorcyclists—who may be in the blind spot of large trucks. Additionally, the system uses both visual and auditory alerts to warn drivers and pedestrians, helping to prevent potential safety accidents.

In addition, the system is equipped with the video monitoring capabilities of a Camera Monitoring System (CMS), utilizing several 1080p high-definition cameras. This setup enables the in-vehicle camera display to provide high-quality video images, assisting the driver in accurately assessing the surroundings of the vehicle. We also offer customized services, allowing us to tailor hardware configurations and software algorithms to meet specific customer requirements.

System Composition

The Ultravision Blind Spot Detection (BSD) system consists of an AI-powered Mobile Digital Video Recorder (MDVR) that supports the BSD algorithm, along with BSD cameras (up to 4, depending on customer needs), a monitor, audible and visual alarm, a speaker, and other related accessories.

Figure 3: Ultravision BSD system product portfolio

Installation Location

Customers have the flexibility to choose the installation locations of each system component based on the specific conditions of their vehicle. The following schematic diagram illustrates a typical installation layout for the COE (Cab Over Engine) model, provided for reference.

Figure 4: Installation location diagram

Implementation Effect

At present, Ultravision BSD system supports VRU monitoring and alarm in eight directions, which can meet the relevant regulatory requirements of the aftermarket, such as CMS, MOIS and BSIS requirements in DVS PSS. To a certain extent, HGV/LGV vehicles equipped with BSD system have the same effect under the relevant EU regulations such as UNECE R151 BSIS, UNECE R159 MOIS, because it does improve the current situation of limited vision of heavy trucks. The effect is slightly different for different models. The following is a schematic diagram of COE model with the cabin on the left.

Figure 5: Schematic diagram of the effect display

In practical applications, the system can accurately identify vulnerable road users—such as pedestrians, cyclists, scooters, and motorcyclists—with an accuracy rate of 96%. It also provides timely voice warnings, particularly when the vehicle is turning, alerting pedestrians and other vulnerable road users to maintain a safe distance.

Figure 6: Schematic diagram of algorithm identification of VRU

Video 1: Audible and visual alarm voice reminder video

Additionally, the system offers an open integration port, enabling seamless connection with third-party fleet management platforms for real-time alarm data uploads. It supports real-time video streaming, GPS location tracking, and, when relevant components are configured, monitoring of tire pressure, fuel usage, and other critical vehicle metrics. This functionality significantly enhances fleet management efficiency.

For BSD systems applied to CBE(Cab-Behind Engines) models, the longer front end of the vehicle results in a larger physical blind spot, which can limit the effectiveness of the BSD camera in monitoring the area directly in front of the vehicle. Currently, the industry is leaning more towards sensor-based solutions to address this issue. This article will not delve into these sensor-based solutions in detail, but if you require more information or have specific needs, please feel free to contact us: Contact Us.

Application Scenarios

This system is commonly used in vehicles with limited direct visibility, such as heavy trucks, trailers, large buses, cold chain vehicles, tankers, and large mining equipment. When these vehicles were first designed and manufactured, they often had long bodies or too high driver seats, which made it impossible for the driver to fully grasp the situation around the vehicle by himself or with the help of rearview mirrors. Additionally, the sheer weight of these vehicles means that any traffic accident involving them can have highly destructive consequences.

As a result, transportation authorities in many countries and regions have introduced regulations mandating the installation of BSD-related safety equipment. For instance, the UK’s TfL DVS 2024: Progressive Safe System (PSS) provides a detailed and technical definition of vehicle blind spots. For more information, please refer to the following article: PSS - Progressive Safe System.

Figure 7: Applicable vehicle models

Innovative applications

Thanks to its technical foundation in visual analysis, this solution is also applied in security scenarios, such as restricted-access cargo monitoring. During transportation and non-loading/unloading phases, it continuously monitors whether unauthorized individuals enter the cargo hold, helping to prevent theft. Below is a case diagram illustrating a related project we have successfully implemented:

Figure 8: Application case diagram

Solution Customization

Ultravsion has our own R&D team, so we can provide customers with more solution flexibility, whether it is different specifications of vehicle recorders and different types of camera hardware combinations, or customized software requirements, and integrates with third-party fleet management platforms. We can provide tailor-made solutions for you. In addition, we also support customized voice alarms in different languages to meet the localization needs of more countries and regions.

Ultravsion is looking forward to receiving your unique needs and product feedback, which is the driving force for promoting our product innovation and advancing the globalization of Vision Zero. Welcome to leave us a message to discuss the feasibility of your project or the after-sales problems you encounter.

For more information about Ultravsion, please visit: www.ulvisions.com

Written by Henry Chia on 5 November 2024.