Of a compact size and with a simple appearance, featuring an integration of functions such as patrolling, picture capturing, air monitoring and leakage detection, and hailed as an “artifact” for management and maintenance of tunnels and underpasses……, the first automatic patrol robot for tunnel & underpass uses in Shanghai has been formally put into use at Siping Road Zhongshan No.2 Road (N) Underpass recently.

　　With growing urban construction in Shanghai, underground facilities such as tunnels and underpasses bring great convenience to people’s daily travels, while management and maintenance of such large infrastructure requires investment of numerous human and material resources and consumes lots of engineering work.

　　Siping Road Zhongshan No.2 Road (N) Underpass is a two-way four-lane road, which is 896m long. According to Wang Jiadong, a staff member at the Municipal and Water Affairs Management Center of Yangpu District, the underpass has a complex underground structure, which, “located underneath Dongzouma Tang, shares the structure with some parts of Rail Transit Line 10, that is, above the roof of the underpass runs Dongzoumatang River while the roof of Line 10 is exactly the bottom plate of the underpass.” In order for effective maintenance patrol, measures such as vehicle patrolling in the daytime and partial blocking of the underpass were taken, which not only faced some operating risks, but also reduced the efficiency of the underpass.

　　Today, when driving in Siping Road Zhongshan No.2 Road (N) Underpass towards Wujiaochang, careful drivers may notice a rectangular machine in orange at the top left of the lane. That machine is the automatic patrol robot for tunnel and underpass uses, which is an unmanned patrol device integrating technologies such as the Internet of Things (IoT) and machine learning, and organically combining a robot body, various sensors and background software. The automatic patrol robot can move autonomously along the guide rail for 24 hours without a break, take pictures of and identify operations inside the underpass, analyze and diagnose abnormalities through artificial intelligence (AI) algorithms and immediately report to background staff.

　　According to Wang Jiadong, the robot is composed of built-in modules such as 3D laser radar, infrared IP dome, capture camera and gas sensor. The 3D laser radar can accurately scan and identify foreign matters on the road within the range of 20cm to 150m and with ranging accuracy of ±2cm; the infrared IP dome is intended for remote monitoring and voice ; the capture camera, triggered by a velocity radar, has functions such as license plate recognition and combination of violation photos and is used to capture violations on lanes; the gas sensor monitors indicators such as temperature, humidity, formaldehyde and CO2 in the air and may serve as the basis of evacuation decisions in emergencies.

　　If the robot body and sensors are “iron feet” and “clairvoyant” of management and maintenance personnel, then background software serves as the assistant and elfin and plays a key role in problem identification, analysis and diagnosis. Feng Jun, a staff member at the underpass maintenance unit, showed the reporter the interface of background software operations on the laptop which he carried with him. Apart from a real-time picture of vehicles in the underpass, the interface also displays the status of alarm information such as loose manhole cover, foreign matters, lamp damages, abnormal parking and tunnel traffic accidents, which information is the diagnosis by background software through AI algorithms.

　　The emergence of an automatic patrol robot has greatly increased the efficiency of management and maintenance. Feng Jun stated that “the automatic patrol robot can increase the patrolling efficiency and monitor the vehicle traffic in the underpass around the clock; if any facility is damaged due to a car accident, the robot can immediately inform us and we will rush to repair the facility in time and broadcast the information on construction through background software to drivers to drive with care; in extreme weathers such as rainstorm and low-temperature rainy and snowy weather conditions, it can help us take corresponding measures in time to prevent ponding, freeze and other hazardous conditions in the underpass; with respect to air quality monitoring, once CO exceeds the limit, we can turn on fans in time to ensure good air quality in the tunnel……”

　　It is worth mentioning that background software of the automatic patrol robot has the capability of machine learning and can constantly identify and analyze traffic conditions and facility operations in the underpass based on existing databases, thus accumulating experience for subsequent video analysis and diagnosis. “For example, once a foreign matter alarm is given by background software, we will whether it is true on the scene and return the result to the background, so that the background can train algorithm models on the basis of such data, thus further increasing the accuracy of diagnosis and getting smarter,” said Feng Jun.

　　After the automatic patrol robot formally comes into use, Construction and Management Commission of Yangpu District will update and upgrade functions of the robot, such as its inspection module, radar identification of foreign matters and camera pixels. Meanwhile, the robot will be combined with the existing BIM management and maintenance system to predict and get to know facility conditions, identify and solve facility faults as early as possible and gradually form a new lifetime maintenance strategy that replaces centralized overhaul with minor and medium repairs.

　　Currently, the automatic patrol robot is installed at only one side of the underpass at the crossing of Siping Road and Zhongshan No.2 Road (N) towards Wujiaochang, which patrols within a range of 300m. Head of the Construction and Management Commission of the District indicated that “by the end of this year, if the automatic patrol robot functions as expected, we will install such robot over the entire length of Siping Road Zhongshan No.2 Road (N) Underpass.”