Safety is the foundation of all urban rail transportation systems, and its importance is paramount. Ensuring that there are no intrusions across railway line boundaries—that no foreign objects enter the safety zone—is key to maintaining safety. In recent years, as rail transportation systems have expanded and passenger numbers have increased, intrusions across railway line boundaries have become ever more frequent.

Examples of track intrusions: Tunnel incursion from drilling rig (left), train collision caused by subway's protective doors (middle), track incursion by adjacent tree (right).

In response, the Office of the Safety Commission of the State Council issued an emergency notice calling for further strengthening of safety measures along urban rail transportation lines, including combating illegal construction work, unauthorized utility connections and building modifications, and intrusions by vehicles, people and foreign objects within protected zones.

However, there are numerous challenges to preventing intrusions across railway line boundaries:

First, due to the operating environment and other factors, train drivers still have to manually identify certain obstacles. However, manual identification has a limited visual range, which presents a safety risk.

Second, driverless train systems are an effective means of resolving this problem, but their core function of autonomous forward environmental awareness is still inadequate, and improvements are urgently needed.

At the same time, as train networks grow more complex, it becomes increasingly difficult to capture the position and status of equipment for each line and platform in real time. In addition, the added impact of unpredictable human factors and dynamic relationships between risk factors mean there are inadequate prediction and early warning capabilities.

Solution：Machine vision with integrated LiDAR

To tackle these issues, Beijing AI for Rail Technology Co., Ltd. (Aiforail) has introduced the Livox Tele-15 LiDAR device equipped with the Train Intelligent Obstacle Detection System (TIDS), which can effectively resolve the aforementioned problem with proactive obstacle detection technology based on environmental awareness. The TIDS system comprises multiple sensors, including a visual camera, Tele-15 LiDAR and millimeter-wave radar.

The Tele-15 was specially created for long-range, high-precision detection. It can sense obstacles within a range of 300m, giving a vehicle ample response time and ensuring safety, while simultaneously being able to precisely sense foreign objects in complex scenarios and perform multi-source heterogeneous data fusion to detect trains, people and obstacles in the rail operations zone.

Tele-15 LiDAR

This detection technology can achieve eyesight protection in the rail operations zone, compensating for the shortcomings of FAO eyesight, sustaining train operations after a signal system is downgraded, reducing wait times, and improving operational efficiency. This technology can also be applied to retrofit existing lines and achieve cross-line operations.

Key advantages of TIDS

Advantages of solution

Cost effective, easy to deploy, strong environmental adaptability

The Train Intelligent Detection System (TIDS) integrates LiDAR capabilities and uses sensing modules composed of a camera and radar. The system can be retrofitted to existing trains at a low cost, is quick to deploy, requires minimal modifications, and does not affect existing vehicle signals and other systems. In addition, the system can function in different time periods, on different lines, in different weather conditions, and in different surrounding environments.

Key technological components of TIDS

The system is easy to deploy and requires minimal modifications

The Tele-15 LiDAR integrates other sensing devices that substantially improve the detection performance of the TIDS system. In terms of detection accuracy, it provides high-precision, centimeter-grade positioning. In terms of the detection range, it can detect vehicles at 300m, pedestrians at 120m and small obstacles at 50m. In addition, detection is not affected by the vehicle's windshield, and the system can provide round-the-clock support in different scenarios, including underground, ground-level and elevated tracks.

TIDS performance parameters

Equipped with radar as the core sensor, the system provides 3 solutions for different types of lines:

The first is a driver assistance system with environmental awareness, which can enhance the driving safety of train drivers;

The second is a backup solution for fully automatic driverless lines when the train signal system is downgraded.

The third is independent sensing vehicle control, which achieves sensing-based safety control without relying on the existing signal system.

Deployment outcomes

Over 1000 units deployed; suitable for a wide range of scenarios

Currently, the TIDS system with integrated LiDAR has been deployed in Beijing, Chengdu, Taiyuan, Hong Kong and other cities on a total of 350 trains for 29 lines, with 1332 units deployed. To date, the system has a total online running time of more than 10 million kilometers.

Image of equipped train for Line 17 of the Beijing Metro

During operations, the system has demonstrated strong adaptability and reliable detection in over 100 complex scenarios, including straight, curved and sloped tracks, as well as in rain, fog, snow and other harsh weather conditions.

The system offers reliable detection in a wide range of different environments

Analysis of operational data for the TIDS system has demonstrated the product's ability to function on different lines, trains, and weather conditions. As a result, it is suitable for all scenarios and offers a cutting-edge solution to the major challenges of driverless train systems.

The system senses the distance to obstacles as the vehicle moves and issue alerts.

SIL4 clearance detection and sensing to stop for vehicles ahead

The Tele-15 radar detection system offers a standardized, data-driven and highly efficient mechanism for improving rail operations and ensuring operational safety. Equipped for a wide range of scenarios, it provides a cutting-edge solution to the challenges associated with driverless rail systems.

Livox Technology Company Limited (Livox) is an independent company founded through DJI's internal incubation program and dedicated to the development of 3D lidar technology.

Livox was founded in 2016 with the resolve to break through technological barriers in the lidar industry and reduce the threshold for mass production by producing high-quality, reliable, and cost-efficient lidar sensors that can be applied in a wide range of fields, including the automotive industry, smart cities, surveying and mapping, and mobile robots.

Over the past few years, Livox has applied for 587 patents for its unique sensor designs and rotating mirror hybrid solid-state lidar technology. Since the Mid product series was rolled out in January 2019, Livox has served more than 1500 customers, from 26 countries and regions, including the United States, Canada, China, Japan, and the European Union.