Every mountain Search and Rescue operation in Hong Kong is a race against time. Each mission typically mobilises more than a hundred personnel from multiple departments, including the police, the fire service, and the Civil Aid Service—and may involve up to two hundred people working around the clock within the “golden 72 hours.”
For rescuers, the hardest moment is not traversing steep ridges or crossing deep valleys, but breaking news to family that their loved one has not survived.
In recent years, the Hong Kong Government has begun to introduce drones to assist in Search and Rescue operations. Chen Benmei (Ben), Professor and Chairman of the Department of Mechanical and Automation Engineering at The Chinese University of Hong Kong and an IEEE fellow, is now collaborating with the police to develop an Unmanned Aerial Vehicle (UAV) system that is not only more precise, but fully automated. “We are so lucky,” he says, “that this collaboration exposes us to real‑world challenges, which have driven us to develop solutions that truly work.” 
With more than two decades of research experience in unmanned systems, Ben joined CUHK from Singapore in 2018. Since then, he has developed a deep affection for Hong Kong’s natural landscapes. From verdant mountain trails to skyscrapers, from ecological conservation to mountain rescue operations, his drones traverse these landscapes. As he explains, building a drone itself is not the hardest part. The real challenge lies in system design—crafting solutions that can adapt to unique demands of each application. In this issue of CUbicZine, we speak with Professor Chen to explore the unexpected possibilities that lie beneath the sky.
The Mountain Search Challenge: Finding Those Who Are Lost
The path that led Ben’s drones into the high‑stakes world began with a chance encounter. During a conservation mission, he met Kevin Laurie, Programme Officer of the IUCN Horseshoe Crab Specialist Group and a retired Senior Superintendent of the Hong Kong Police Force. That meeting would eventually bring drone technology out of the laboratory and into real life‑and‑death scenarios. Recalling his first impression of the team’s drone demonstration, Kevin says, “What most impressed me was the fusion of artificial intelligence with drones.” Through Kevin’s network, in 2025 Professor Chen and Research Assistant Professor Chen Xi led their research team in launching a five‑year collaboration with the Kowloon East Police Region. The project focuses on developing a digital‑twin integration system, combining unmanned systems with AI large language models, to support Search and Rescue operations.
This year, the team conducted a series of field demonstrations in forested areas in Hong Kong. The challenges were formidable: vast expanses of woodland, complex terrain, and missing persons who often shelter in dark, heavily shaded areas. To address these difficulties, the team is working on integrating visual imaging, communication signals, Bluetooth devices, and base‑station systems, refining the technology through constant dialogue with end users—the police—and incremental improvements during each demonstration.
In the first phase, the system relied on mobile phone signals for positioning, achieving an effective detection range of 500 to 700 metres. At the same time, aerial imagery processed through backend software reconstructs the surrounding environment into an instant 3D model, allowing command personnel to adjust search strategies in real time. Meanwhile, consideration is being given to equipping search teams with smartwatches, to transmit their movement paths in real time, enabling more efficient coordination and deployment.
The next phase will push even further. Fully autonomous drones are being developed to integrate sensors capable of detecting targets that emit no mobile phone signals at all. Once the system is fully calibrated, Professor Chen notes, the drones may even be able to fly through dense undergrowth for close‑range scans—greatly enhancing search effectiveness in shaded areas. Field tests have already been conducted in multiple scenarios such as caves and tunnels.
Over the past year, Kevin has been responsible for liaising and coordinating various test flight locations. With a laugh, Professor Chen remarks that this “super-connector” has by now become “half an engineering expert.”
Academia Meets the Police: From Jargon to Reality
Having participated in numerous technology collaborations during his police career, Kevin admits that many projects were “over-promising”. What sets Professor Chen’s team apart, he says, is their ability to deliver. “I gave them problems and they came back with solutions.”
Still, acting as a communication intermediary between academia and the police is no easy task. Kevin describes every conversation as a nightmare – for instance, when the professor’s team speaks of “digital twins” to the police or requests testing of SLAM (Simultaneous Localisation and Mapping) algorithms in tunnels. “Why can’t you just call it a virtual replica? Or simply say we need to test in a tunnel?” Over time, he has built his own internal translation guide, anchored by a single rule: the simpler, the better—because the police must be able to use the technology instantly under extreme conditions. Only by effectively bridging the gap between technical terminology and police jargon can both parties’ needs and expectations be met, enabling innovation to move out of the lab and into real‑world use. He believes, “This project will change the future. And unless something better comes along, it will be functional forever.”
Although the project remains in the research phase, Professor Chen Xi notes that some of the technologies are expected to be deployed in real operations within a year, with full implementation anticipated within three to five years. It is estimated that this will reduce operation costs by more than 50%. The true value of drones lies in their ability to operate where humans cannot—or should not—go. In recent years, the team’s unmanned systems have been tested with government departments in Hong Kong and Shenzhen across a range of high‑risk, labour‑intensive scenarios, including mountain and maritime rescue, ecological and environmental monitoring, powerline and pipeline inspections, and intelligent building assessments. Ben pointed out that drones actually have a very wide range of applications; for example, the framework for searching for missing persons is largely the same as that for searching for juvenile horseshoe crabs.
Saving Living Fossils with Drones and AI
Along the shores of Ha Pak Nai, faint yet unmistakable tracks trace the quiet presence of a creature older than the dinosaurs themselves—the horseshoe crab. Having survived for more than 400 million years, this ancient life form now teeters on the edge of extinction. In recent years, the number of juvenile horseshoe crabs in local waters has dropped sharply, and the Tri-spine Horseshoe Crab has even been classified as an “Endangered” species. In 2023, Ben led a research team participating in the Ocean Park Conservation Foundation’s Horseshoe Crab Population Survey, marking the first time drones and artificial intelligence were introduced to track the movements and habitat distribution of juvenile horseshoe crabs.
Detecting these animals, Professor Chen explains, is no easy task. Small and cold‑blooded, they blend seamlessly into the mudflats, using natural camouflage to evade notice. Learning from Kevin, the horseshoe crab expert, that the juvenile horseshoe crab leaves distinctive trails on the sandy mud, the team devised an innovative approach that turned behaviour into signal. By combining RGB imagery with AI‑based tracking, the system captures the brief moments when the juvenile crabs emerge onto the mudflats to forage. The result achieved an accuracy rate of over 90 percent—greatly reducing manpower needs while eliminating the need for researchers to tread across fragile mudflats, thereby minimising disruption to the ecosystem.
For Professor Chen, what makes this project especially meaningful is its collaborative nature. “Collaborating in the field is the most rewarding part. We don’t know the lifestyle habits of horseshoe crabs; only conservation experts do.” He stresses that the project’s scientific value lies not simply in locating the animals, but in transforming collected data into actionable insights that can drive effective conservation strategies and help secure a future for this ancient species. The project was later supported by the Environment and Conservation Fund.
Exploring Drone Markets in the Concrete Jungle
To drive their research into real‑world applications, the team establishedCU-Craft Limited in 2023. The company integrates unmanned systems with AI-powered inspection solutions to meet the growing demand for infrastructure and building inspections in Hong Kong’s densely packed urban landscape.Current inspection methods rely on manual work in confined spaces, are dangerous and labor‑intensive, and the data collected is scattered across different systems, making it hard to use in real time.
As industrial expectations for technology continue to rise, Professor Chen Xi believes the team’s expertise places them at a significant advantage. “Traditional drones rely on satellite positioning,” he explains. “But in Hong Kong’s concrete jungle, signals are often unstable when inspecting high-rise buildings, making it difficult for drones to navigate freely. We are focused on developing fully autonomous inspection technology that operates without satellite signals.” CU‑Craft specialises in an automated building structural inspection and management platform, which assesses conditions both inside and outside buildings to assist with property maintenance and repairs. Professor Chen Xi estimates that the technology has the potential to reduce inspection costs by approximately one-third.CU‑Craft successfully entered the market by securing contracts with five government clients across Hong Kong, laying the groundwork for further expansion into the commercial market.
In addition to the two professors, the research group consists of around twenty PhD students, postdoctoral fellows, and researchers from Hong Kong, mainland China, and overseas. Over time, the team has built long‑term academic partnerships with universities across the mainland, including institutions in Beijing, Shanghai, Harbin, Fuzhou, Xiamen, and Guangzhou.
The Core of Unmanned Aircraft
Whenever Ben finds a rare opportunity to return to the mountains and forests, he looks up with quiet satisfaction as drones glide through training exercises. To most observers, they are simply metal birds drifting freely in the air. Yet in the mind of this expert, each flight represents a carefully solved equation. “How to fly is a science problem. Every flight is not a random journey. Every path is calculated according to specific requirements.” Under the same sky, a single drone in motion and a coordinated fleet of drones constitute entirely different technical systems. Likewise, aerial photography for 3D environment reconstruction relies on a separate set of complex software engineering.
In his view, drones have never belonged to a single discipline. Instead, they embody the convergence of many fields—mechanical design, electronics, software engineering, sensing technologies, and real‑world applications. “We talk a lot about so-called unmanned drones. But in truth, it is the man behind them who is the most important. Technology is created by humans.”
Scholar Encyclopedia:
Technology Start-up Support Scheme for Universities (TSSSU)
The establishment and growth of CU-Craft Limited have been supported by TSSSU. Professor Chen Xi, the company’s director, candidly remarked: “CU-Craft is the first company we have put into operation. The funding provided by TSSSU played a critical part in the establishment and operation of the company.”
Launched by the Innovation and Technology Commission in 2014, TSSSU was designed to support research teams from local universities in establishing start‑ups. The scheme covers essential expenses required for company formation and early-stage operations, thereby promoting the commercialisation of research outcomes. From the 2023/24 financial year onwards, an “Extension Scheme” was introduced, offering a one-to-one matching funds alongside private investment to further accelerate the growth of university‑based start‑ups.
InnoHK – the holistic R&D Hub
Professors Chen Benmei and Chen Xi joined the CUHK’s Hong Kong Centre for Logistics Robotics (HKCLR) in 2020, where their drone research team has since grown to include more than twenty researchers.
HKCLR is funded under the InnoHKof the Innovation and Technology Commission. According to Professor Chen Xi, InnoHK support has enabled the team not only to expand its research capacity through talent recruitment, but also to strengthen industry engagement and advance market development.
InnoHK is a key initiative of the Hong Kong Government, aimed at developing Hong Kong into a global hub for scientific research collaboration. By fostering close partnerships among universities, research institutions, and industry, the programme promotes collaborative research and development across disciplines.At present, InnoHK comprises three major R&D platforms: Health@InnoHK (health technology), AIR@InnoHK (artificial intelligence and robotics), and SEAM@InnoHK (sustainability, energy, advanced manufacturing and materials).
More
